property rights theory and ownership of firm …
TRANSCRIPT
PROPERTY RIGHTS THEORY AND OWNERSHIP OF FIRM-SPECIFIC
ADVANTAGES: THE IMPLICATIONS OF CONTRACTING AND LICENSING
WITHIN THE MULTINATIONAL FIRM
by
Catherine Magelssen
A Dissertation submitted to the Graduate School-Newark
Rutgers, The State University of New Jersey
in partial fulfillment of the requirements
For the degree of
Doctor of Philosophy
Graduate Program in Management
written under the direction of
Susan Feinberg
and approved by
____________________________
____________________________
____________________________
____________________________
Newark, New Jersey
October, 2014
© 2014
Catherine Magelssen
ALL RIGHTS RESERVED
ii
ABSTRACT OF THE DISSERTATION
Property Rights Theory and Ownership of Firm-Specific Advantages: The
Implications of Contracting and Licensing within the Multinational Firm
By Catherine Magelssen
Dissertation Director:
Professor Susan Feinberg
Firm-specific advantages (FSAs) play a critical role in the theory of the
multinational firm. Firms establish foreign operations when their FSAs are not suitable
to outsource or license in the market. However, the same assets deemed unsuitable for
external contracting and licensing are extensively contracted and licensed within the
multinational firm. The parent and/or subsidiaries that are the economic owners of the
assets contract other entities within the firm to perform activities such as R&D,
manufacturing, and distribution and pay them a guaranteed return for their activities.
Internal ownership of FSAs has implications on the risks borne, incentives, resource
allocation, and power distribution within the firm. Using a unique, confidential dataset
on the internal transactions of multinational firms, including intra-firm product flows,
economic ownership of FSAs, financials, and detailed specifications of the activities of
subsidiaries within the firm, I examine the determinants of the structure of economic
ownership of FSAs as well as the impact of FSA ownership on innovation within the
multinational firm.
iii
Dedication
To my husband, Hans, and to my family. Thank you for your love,
encouragement, help, and support.
iv
Acknowledgements
I would like to express my sincere gratitude to my dissertation chair, Susan
Feinberg for her constructive advice, persistence, and support. I am also grateful for my
dissertation committee members Fariborz Damanpour, Michelle Gittelman, and Tom
Prusa. They provided invaluable advice and feedback and were a continuous source of
encouragement. I would like to acknowledge Xavier Martin and Rajneesh Narula, who
provided comments at the AIB doctoral consortium, and the participants at seminars for
their thoughtful suggestions. I am grateful for all of the friends that I have met in the
doctoral program, and appreciate their encouragement and feedback. I would also like to
express my thankfulness towards Rutgers Business School for its financial support.
I am deeply indebted to the consulting firm and experts, whose support made this
dissertation possible.
v
Table of Contents
1. Introduction ................................................................................................ 1
2. Theoretical Background ............................................................................. 6
FSAs and the MNC .................................................................................... 6 2.1
Property Rights Theory and the Firm ......................................................... 7 2.2
3. Data .......................................................................................................... 16
Coding of Data ......................................................................................... 18 3.1
FSA Ownership Structures ....................................................................... 22 3.2
MNC Entity-Level Data ........................................................................... 36 3.3
Summary .................................................................................................. 41 3.4
4. Inside the MNC: Structuring Ownership of Firm-Specific Advantages . 42
Introduction .............................................................................................. 43 4.1
FSAs and FSA Ownership Structures ...................................................... 48 4.2
Theoretical Background ........................................................................... 50 4.3
Hypotheses ............................................................................................... 57 4.4
Methods .................................................................................................... 59 4.5
Results ...................................................................................................... 70 4.6
Discussion ................................................................................................ 80 4.7
5. Subsidiary Ownership of FSAs and Innovation ....................................... 88
Introduction .............................................................................................. 88 5.1
Contextual Background ............................................................................ 93 5.2
Theoretical Background ........................................................................... 96 5.3
Hypotheses ............................................................................................... 99 5.4
Methods .................................................................................................. 104 5.5
Results .................................................................................................... 116 5.6
Discussion .............................................................................................. 124 5.7
6. References .............................................................................................. 128
vi
List of Tables
Table 1: Coding of FSAs .................................................................................................. 19
Table 2: Coding of FSA Ownership Structures ................................................................ 24
Table 3: Mean MNC Characteristics and FSAs by FSA Ownership Structure ................ 35
Table 4: Subsidiary Descriptive Statistics by Country ..................................................... 37
Table 5: Mean Value of Subsidiary Characteristics by FSA Ownership Type ................ 40
Table 6: Characteristics of FSA Ownership Structures .................................................... 49
Table 7: FSA Contractibility ............................................................................................. 57
Table 8: Contractibility Dimension Word Counts ............................................................ 65
Table 9: Mean MNC Characteristics and FSAs by FSA Ownership Structure ................ 70
Table 10: Descriptive Statistics and Correlations ............................................................. 71
Table 11: Probit and Bivariate Probit Results Predicting Shared FSA Ownership .......... 75
Table 12: Multinomial Logit Results ................................................................................ 78
Table 13: Breakout of FSA Ownership by Subsidiary Type ............................................ 94
Table 14: Balancing Test for Propensity Score Matched Sample .................................. 113
Table 15: Balancing Test for Change in FSA Ownership Matched Sample .................. 115
Table 16: Descriptive Statistics and Correlations for Initial Conditions Analysis ......... 116
Table 17: Descriptive Statistics and Correlations for Propensity Score Analysis .......... 117
Table 18: Subsidiary FSA Ownership Change ............................................................... 118
Table 19: Random Effects Negative Binomial Results .................................................. 120
Table 20: Results for Transferring Ownership of FSAs Away from Subsidiary ............ 123
vii
List of Figures
Figure 1: Example of FSA Ownership and Contracting and Licensing Arrangements .... 10
Figure 2: Types of FSAs ................................................................................................... 21
Figure 3: Sole Ownership Structure .................................................................................. 25
Figure 4: Shared Ownership Structure .............................................................................. 27
Figure 5: Separate Ownership Structure ........................................................................... 29
Figure 6: Mixed Ownership Structure .............................................................................. 31
Figure 7: FSA Ownership Structure by Industry .............................................................. 34
Figure 8: Number of Changes in FSA Ownership by Year .............................................. 39
Figure 9: Ownership Structure by Type ............................................................................ 72
Figure 10: Percentage of FSA Types Shared vs. Separate................................................ 73
Figure 11: Average Patenting and Citations of Patents for FSA Ownership Change .... 119
1
1. Introduction
The ownership of firm-specific advantages (FSA) is central to the theory of the
multinational firm (Dunning, 1977) and firms in general (Drucker, 1995; Grant, 1996;
Kogut and Zander, 1992; Penrose, 1959). FSAs, also referred to as knowledge assets and
competencies in the management literature, are the proprietary assets that provide the
firm with a competitive advantage (e.g. Rugman and Verbeke, 2001). Firms establish
foreign operations in order to capitalize on the ownership of firm-specific advantages,
which can be transferred across the firm at a relatively low cost (Hymer, 1960; Caves,
1971). Although the theory of the multinational firm (MNC) suggests that firms establish
foreign subsidiaries when their FSAs are not suitable to contract or license in foreign
markets, MNCs contract and license the same FSAs within the firm to their subsidiaries
in those same foreign markets. The parent and/or subsidiaries (entities) that own the
FSAs (FSA owners) internally contract other entities (FSA users) within the firm to
perform activities such as research and development, manufacturing, and distribution and
pay the FSA users a guaranteed return for their activities. The FSA ownership structure
has implications on the risks borne by the various affiliates, incentives, resource
allocation, and power distribution within the MNC. While many researchers have studied
the external licensing and contracting relationships of the MNC, the internal licensing
and contracting relationships amongst affiliates have thus far been unexplored. The lack
of research in this area is no doubt due to the lack of publicly available data.
This dissertation contributes to the international business literature by examining
three questions related to MNC internal FSA ownership (or simply “FSA ownership”).
First, how do MNCs internally organize ownership of their most important value-
2
generating assets? Second, how do FSA characteristics affect MNC choice of FSA
ownership structure? Third, how does FSA ownership affect affiliate innovation? I also
explore the role of tax haven FSA ownership.
A natural question arises as to whether the internal structure of FSA ownership is
simply an artifact of tax avoidance. Recent U.S. Senate hearings on Apple and other
multinational companies, and U.K. Parliamentary hearings on Starbucks, Amazon, and
Google, have captured the public’s attention and highlighted the role of shifting
ownership of FSAs offshore as a means of avoiding taxes. While some companies have
tax haven affiliates that perform research and development, manufacturing, or
distribution activities, others have tax havens that are no more than a mailbox (Drucker,
2010). Clearly, tax avoidance plays a role in FSA ownership structures. However, very
little is known about MNC FSA internal ownership outside of what is reported in the
media due to alleged tax avoidance.
Transfer pricing research focuses on predicting the appropriate transfer price of
goods within a MNC (e.g. Tsay, 1999) and examines whether MNCs shift profits from
high to low tax jurisdictions (Grubert and Mutti, 1991; Grubert, 2003; Mutti and Grubert,
2008; Dischinger and Riedel, 2010). Recent research recognizes that some MNCs
engage in extensive profit shifting whereas others do not (Grubert, 2003; Overesch and
Schreiber, 2008). The research on transfer pricing examines where profits are shifted, it
does not examine the differences in MNC internal FSA ownership structures. It also fails
to examine how the ownership of FSAs influences MNC real operations.
Chapter 2 provides the theoretical background for understanding the internal
contracting and licensing of FSAs that occurs within the MNC. Consistent with resource
3
and knowledge-based perspectives of the firm (e.g. Barney, 1991; Amit and Schoemaker,
1993; Grant, 1996; Kogut and Zander, 1992), the firm is viewed as a heterogeneous
bundle of assets. Firm-specific advantages are considered the most important assets for
the firm. I build on property rights theory, which assumes that a set of rights can be
attributed to each asset or resource and lays the groundwork for predictions on the
allocation of those rights to parties in an exchange relationship. The MNC has an internal
network of exchange relationships. I argue that property rights theory compliments the
resource based perspective by enabling predictions on how the rights to the MNC’s key
value generating assets are allocated across its network of subsidiaries.
Chapter 3 provides a contextual background and an overview of the data.
Particularly since the dataset is new and relatively little is known about MNC internal
ownership of FSAs, it is worth providing some descriptive statistics. To conduct my
research, I assembled a unique confidential panel dataset obtained from a consulting firm
that advises MNCs on transfer prices. My dataset consists of the intra-firm transactions
of 102 MNCs and their parent and subsidiaries in an unbalanced, panel dataset from 1997
to 2012. The dataset includes detailed data on the economic ownership of FSAs within
the firm, contracts between the FSA owners and FSA users that clearly delineate the
rights and responsibilities of each party, mergers and acquisitions (M&As), changes in
ownership structure, tax haven ownership, financials, and product flows. I combined this
data with data from Bureau Van Dijk’s Orbis database, Thomson Financial M&A
database, and United States Patent Trademark Office. I used the combined data to
construct multinational firm-level and subsidiary-level datasets. MNCs vary by whether
FSA ownership is centralized into one entity or dispersed across many entities. Similar
4
to the types of ownership structures that emerge in markets, I identify four different types
of FSA ownership structures that MNCs use: 1) sole ownership, 2) shared ownership, 3)
separate ownership, and 4) mixed ownership. I examine descriptive statistics associated
with each structure and with tax haven ownership.
Chapter 4 builds on property rights theory to understand how MNCs allocate the
economic ownership of FSAs to different units within the firm. This chapter investigates
the FSA characteristics that influence the choice of ownership structure. Property rights
theory suggests that in an exchange relationship, the party whose activity makes the
largest contribution to the creation and maintenance of the asset should own the asset.
However, MNCs must balance these considerations against other potential advantages
such as reduced administrative and monitoring costs, reduced bargaining problems, tax
minimization, and greater internal knowledge sharing. This chapter argues that MNCs
with knowledge-intensive, tacit knowledge assets are more likely to have shared FSA
ownership structures, whereas MNCs with independent FSAs and FSAs that require
local, downstream inputs are more likely to have separate FSA ownership structures. The
empirical results strongly support property rights theory predictions.
Chapter 5 investigates whether and to what extent FSA ownership affects
innovation. Although FSAs have long held a central role in the theory of the MNC, the
effects of internal ownership on innovation remains unexamined. Innovation is
inherently difficult to monitor and control. Property rights theory suggests that
ownership provides two incentives for investing in the creation of the asset: 1) the ability
to appropriate income from the innovations created, and 2) the ability to control both the
asset and its future direction of development. I find that subsidiary FSA ownership is
5
positively associated with innovation, and that transferring ownership away from the
subsidiaries that create the FSAs has a negative effect on innovation.
This research makes several contributions. First, I contribute to the theory of the
MNC by shedding new light on how FSAs are internally organized, developed, and
managed within MNCs. Due to the lack of publicly available data, previous research has
not been able to open up the black box of internal MNC transactions. FSA ownership is
important because through the contractual relationships risk is shifted from the FSA users
to the FSA owners, and the FSA owners are centrally positioned in the firm’s internal
network of financial, knowledge, and product flows. Therefore the FSA owners can have
a significant effect on MNC investment and innovation. Second, I extend property rights
theory to inside the firm and identify the types of FSAs that are more likely to be solely
owned by one entity, co-owned by two or more entities, or separately owned by different
entities within the MNC. This research suggests that internalization is not sufficient for
resolving the problems associated with contracting for knowledge and provides insight
into the ability of property rights theory to explain internal firm behavior. Third, I
investigate how the internal governance of FSAs affects the creation of new FSAs by
analyzing how subsidiary FSA ownership influences subsidiary technological innovation.
Finally, I explore the role of tax havens and FSA ownership.
6
2. Theoretical Background
FSAs and the MNC 2.1
The management of firm-specific advantages (FSA) is at the core of theorizing
about the existence of MNCs and firms in general (Buckley and Casson, 1976; Dunning,
1977; Rugman, 1980; Barney, 1991; Grant, 1996; Kogut and Zander, 1992). The firm is
conceptualized as a bundle of heterogeneous resources and assets (Amit and Shoemaker,
1993; Penrose, 1959), of which firm-specific advantages are considered the most
important (Barney, 1991; Rugman and Verbeke, 2001). FSAs provide firms with
competitive advantages, enable them to generate profits, and expand abroad. FSAs are
the reason why multinational firms exist. The literature on MNCs suggests that firms
internalize transactions when FSAs are not suitable to contract or license in the market
(Buckley and Casson, 1976; Rugman, 1980). Market failures and uncertainties create
bargaining problems and agency costs, which make it more efficient to transact inside the
boundaries of the firm than in markets. However, these same FSAs deemed unsuitable
for external contracting and licensing are extensively contracted and licensed between
subsidiaries and/or the parent within the MNC.
Referred to as competencies, knowledge assets, and firm-specific advantages in the
management and international business literature, FSAs are considered by many as the
most important source of above normal returns (e.g. Rugman and Verbeke, 2001;
Drucker, 1995; Spender and Grant, 1996). FSAs include technologies, patents, brands,
know-how, and organizational routines (Birkinshaw, Nobel and Riddensdale, 2002).
Rugman and Verbeke (2001) explain that FSAs are the unique company strengths, which
include a broad range of functional, technological, and organizational proprietary assets
7
and know-how. FSAs include technological and/or marketing knowledge, and superior
managerial capabilities to control and coordinate international transactions (Buckley and
Hashai, 2009). FSAs provide the firm with a competitive advantage and are thus a key
source of above normal profits.
This research applies property rights theory (e.g. Grossman and Hart, 1986; Hart
and Moore, 1990) to the internal structure of the MNC. Although property rights theory
is typically applied to external exchange relationships, extending it to the internal
contracting structure and relationships between entities within the MNC can enhance our
understanding of the firm. In contrast to transaction cost economics which focuses on the
characteristics of transactions, property rights theory focuses on the characteristics of
assets and asserts that ownership of an asset can be broken into sets of rights to the asset.
Taking the view that the firm is a bundle of heterogeneous assets, property rights theory
allows us to ask how the rights to the MNC’s key value-driving assets are allocated
across its network of subsidiaries.
Property Rights Theory and the Firm 2.2
The Theory of the Firm (Coase, 1937) spawned a large literature on the
internalization of transactions within the boundaries of the firm. One stream of research
that developed from Coase (1937) is property rights theory (Grossman and Hart, 1986;
Hart and Moore, 1990). Property rights theory focuses on the question of who owns the
property rights of the assets in an exchange relationship (Foss and Foss, 2001). Property
rights are defined as “any sanctioned behavioral relations among decision makers in the
use of potentially valuable resources; such sanctioned behaviors allow people the right to
use resources within the class of non-prohibited uses” (Asher et al, 2005: pp. 7).
8
Property rights may include the right to use, exclude others from use, appropriate income,
and transfer or invest in the resource (Foss and Foss, 2001). The multidimensionality of
rights to an asset means that different entities can hold different rights to the asset. In the
case of the MNC, this means that different subsidiaries can hold different rights to the
MNC’s key assets. For example, a subsidiary may have the right to manufacture a
product, but not the right to sell the product in particular markets. Viewing a firm as a
bundle of assets, the rights of subsidiaries to the MNC’s FSAs can affect their role,
appropriation of income, incentives, and formal linkages to other subsidiaries within the
firm (operational structure).
A fundamental assumption in property rights theory is that contracts are
incomplete - it is impossible to specify all terms and contingencies in advance.
Unforeseen contingencies and the costs of writing and enforcing contracts make contracts
incomplete (Tirole, 1999). Internal transactions of goods, services, and knowledge assets
between MNC entities face similar challenges as external market transactions. FSAs, by
nature, are incomplete. Foss and Foss (1998) claim that intangible assets are a source of
incomplete contracts that lead to imperfectly specified rights. Moreover, MNCs are faced
with diverse risks and uncertainties from their global operations. Shifts in global
demand, prices, and costs create large risks within the MNC causing incomplete internal
contracts.
Incompleteness gives rise to two types of entities in an exchange relationship: 1)
specific rights owner, and 2) residual rights owner. The specific rights owner performs
activities as specified in a contract in return for a guaranteed income. Because the
specific rights owner is guaranteed a return, its exposure to risk is mitigated. The
9
residual rights owner bears the risk associated with the exchange relationship. It receives
the profit or loss based on the difference between the total income and the payments
promised to the specific rights owner (Jensen, 1986). The classical example of a specific-
residual relationship is the employment contract where individuals prefer to be hired as
employees directed to perform tasks in return for a guaranteed wage (Coase, 1937). The
entrepreneur (residual rights owner) takes the residual, fluctuating income (or loss) above
the amount promised to the employees (Coase, 1937).
Extending this concept to MNC and the contractual relationships between entities,
the FSA owners within the MNC are the residual rights owners. The affiliates that
license the FSAs, or are contracted to perform various functional activities (FSA users),
are the specific rights owners. The FSA owners contract the FSA users within the firm to
perform various functions in return for a guaranteed return. The FSA owners bear the
risk of the fluctuating income. By allocating various rights between the related affiliates,
MNCs are able to shift the effects of risk and uncertainties to various units within the
firm.
The concept of residual ownership is equivalent to economic ownership (Barzel,
1997). Economic ownership, which is conceptually distinct from equity and legal
ownership (Barzel, 1997), is based on ultimately bearing the costs and risks associated
with the activities. Fama and Jensen (1983) note that most organizational forms have
contract structures that limit the risks undertaken by most agents through specifying fixed
10
or incentive payoffs.1 The following provides an example of residual-specific rights
ownership within the MNC.
Figure 1: Example of FSA Ownership and Contracting and Licensing
Arrangements
The FSA owners are the entities that bear the risk and not necessarily the entities
that physically create the MNC FSAs. For example, a pharmaceutical MNC has an R&D
affiliate located in India. The FSA owner, which is a U.S. affiliate, contracts the Indian
affiliate to perform R&D activities. The FSA owner agrees to reimburse the Indian R&D
affiliate for all of its R&D costs, plus a 15% mark-up on its R&D costs for the R&D
1 While economic ownership may overlap with equity and legal ownership, it is conceptually distinct.
Alchian and Demsetz (1973) explain that property rights deal with the right to use the resource, and not
legal ownership per se. Barzel (1997) distinguishes between economic and legal rights, viewing economic
rights as the more relevant concept of property rights. Barzel (1997) argues that economic rights, or the
right to residual income, are the ends that agents seek whereas legal rights are the means to achieve the
ends. While legal rights are recognized by governments and enhance economic rights, they are neither
necessary nor sufficient for economic rights (Barzel, 1997). Under property rights theory, the concept of
residual rights ownership can be detached from equity ownership (Fama and Jensen, 1983). Zingales
(2000) argues that viewing the shareholders of the firm as residual claimants is a narrow view of the
residual claimant. Fama (1980) goes so far as to suggest that ownership of the firm is “an irrelevant
concept” (p 290). The view held in this paper is that, within the MNC, the economic owners of the MNC
FSAs are the residual claimants within the MNC.
FSA Owner Drug Discovery,
Drug Formulas, and
Trademarks
UK licenses FSAs from U.S.
U.S. Subsidiary
U.S. contracts India to perform R&D
UK Subsidiary
Sales and
Distribution
Shifts
Development
Risk India Subsidiary
Research and
Development
11
services. The Indian affiliate is a specific rights owner as it is guaranteed to earn a return
(markup on its R&D expenses), regardless of the outcome of its activities. If the R&D
affiliate is not able to successfully develop a new drug, or if it takes it an extra five years
to do so, the U.S. affiliate is responsible for the losses associated with the Indian
affiliate’s R&D activities and the Indian R&D affiliate continues to earn the stable 15%
return on its R&D activities. If, however, the Indian affiliate successfully develops a new
product, the U.S. affiliate, as the FSA owner, receives any profit that is above the 15%
return given to the Indian R&D affiliate. In other words, it receives the above normal
returns (residual income) for any newly created FSAs.
Thus, the entities within the MNC can be categorized into two groups: 1) FSA
owners, and 2) FSA users. The FSA owners can be the parent and/or affiliates within the
MNC. Some FSA owners may perform operational activities, such as R&D,
manufacturing, or distribution, whereas others may be tax haven entities with no
operational activities. MNCs may have one, several, or many FSA owners. FSA users
are the affiliates and/or parent that do not own the economic rights to the MNC FSAs.
These entities are either contracted by the FSA owner(s) to perform specific activities
such as R&D, manufacturing or distribution, or they license the rights to use the MNC
FSAs from the FSA owner.
The concept of residual income is equivalent to the strategy concept of above
average returns. Residual income is the income in excess of normal (market) returns for
the activity performed. Above normal returns are attributed to firm-specific advantages
(e.g. Rugman and Verbeke, 2001). Thus, FSAs are the important value-drivers of
12
multinational firms. Firm-specific advantages are dynamic; firms must continuously
work to develop and maintain their competitive advantage (Buckley and Casson, 1976).
The development of FSAs is risky. The separation of specific rights from residual
rights facilitates the risk sharing between the agents in an exchange and the allocation of
economic rents. It is important to note that there are downsides to FSA ownership. Fama
(1980) states that the asset owner, as residual claimant, receives the “uncertain and
possibly negative difference between total revenues and costs at the end of each
production period” (pp. 290). The residual rights owner (Alchian and Demsetz, 1972) is
also called residual claimant (Klein, 1983), residual risk bearer (Fama, 1980; Fama and
Jensen, 1983), and economic rights owner (Barzel, 1997) in the literature. In terms of the
MNC, the affiliates that take on the risks of developing FSAs are the residual rights
owners. These entities bear the consequence of failure to achieve above normal returns.
The role of the residual rights owners is to insulate the other entities within the
MNC from risk. Through guaranteeing FSA users a specified return on their activities,
risk is shifted across globally dispersed locations. This has several important
implications. First, it determines the risk exposure of the MNC entities. Different
contractual terms may shift different levels of risks to or from the FSA users and FSA
owners. Second, FSA ownership determines the entities’ investment incentives. While
FSA owners are motivated to make investments that increase the value of the FSAs, non-
owners have reduced incentives for investment. As a result, the structure affects
economic behavior and outcomes (Kim and Mahoney, 2005). Third, it determines which
13
entities control the financial resources of the firm.2 In return for bearing the risk, the FSA
owners are entitled to the residual income from the FSAs. This empowers the FSA
owners in resource allocation. Property rights theorists note that asset ownership
provides the owner with power and control over the operations (Hart and Moore 1990;
Rajan and Zingales, 2000). The residual rights owners have the ability to coordinate and
allocate tasks to the non-owning users and the power to punish the users by withholding
resource allocations, re-directing tasks, and exiting businesses (Rajan and Zingales,
2000). Grossman and Hart (1986) argue that residual rights ownership is important
because the owner can influence the solutions to problems and the strategic direction of
future intangible assets. As a result, I expect that FSA ownership will have a significant
effect on MNC incentives, resource allocation, and strategic decisions.
The relationship between the FSA owners and FSA users within the firm is aligned
with Teece’s (1986) observation of the need to own complimentary activities in order to
capture the returns from innovation. The purpose of property rights is to maximize value
and transfer resources to their best uses. Different ownership structures emerge in
response to “economic problem of allocating scarce resources” (Kim and Mahoney,
2005). Allocation of property rights is viewed as an efficient means of dividing
economic rents and avoiding inefficient expropriation and underinvestment (Asher et al,
2005; Rajan and Zingales, 2000). Property rights facilitate implementing value-creating
activities so that resources are channeled to high yield uses (Kim and Mahoney, 2002).
The contractual relationships with FSA users prevent bargaining problems within the
MNC over which entities control the FSAs and the allocation of income from the FSAs
2 Although the parent can repatriate income, tax consequences and the bargaining power of the FSA owner
may discourage repatriation.
14
within the firm. It also prevents future conflicts as to whether an affiliate created a
particular MNC FSA. By contracting the FSA users within the firm to perform various
functional activities, the FSA owner is entitled to the ownership of any FSAs that the
FSA user creates while under the contractual relationship. Therefore the FSA owner will
continue to own any future developed FSAs.
From the ownership of FSAs emerges the internal MNC network of exchange
relationships. As noted by Grant (1996) the drive towards specialization within the firm
creates a need for coordination mechanisms. In markets, the coordinator is the
entrepreneur. Within the MNC, it is the FSA owners which contract and license the FSA
user entities. Through contracting and licensing affiliates, the FSA owners coordinate
MNC activities.
The internal ownership of FSAs is important for several reasons. First, ownership
of FSAs is central to the theory of the multinational firm. Although an extensive amount
of research has been conducted on FSA ownership at the MNC-level, FSA ownership
within the MNC is unexplored. Through exploring questions such as how internal
contracting and licensing is similar to or distinct from external market contracting and
licensing, we can advance the theory of the multinational firm. Second, research on
internal FSA ownership can deepen our understanding of the internal organization and
network structure of the firm (e.g. Bartlett and Ghoshal, 1989; Gupta and Govindarajan,
1991). FSA owners, as the entities that contract and license FSAs to other entities within
the MNC, are centrally positioned in the MNC internal network of financial, knowledge,
and product flows. Third, internal licensing and contracting separates ownership from
control of real operations. The economics literature suggests that various adverse
15
incentive effects arise from the separation of ownership and control. It is important to
understand its impact on organizational outcomes. Fourth, FSA ownership affects the
power distribution and decision making authority within the MNC. The FSA owners are
legally entitled to the residual profits of the MNC (e.g. Internal Revenue Code Section
1.482; OECD Guidelines).3 Once established, transferring FSA ownership can be
prohibitively costly. Their contracting and licensing activities combined with entitlement
to profits means that FSA owners may play a significant role in determining MNC
resource allocation and strategy. Research on affiliate-level FSA ownership has the
potential to make many contributions to the international business literature.
3 OECD Guidelines as well as local country regulations require that entities pay for the FSAs that they
acquire from other entities within the MNC.
16
3. Data
In order to test the predictions, I compiled a novel dataset of MNC FSA
ownership. The primary source of data was MNC transfer pricing reports from a
consulting firm. Transfer pricing is the intercompany pricing of goods, services, and
intangible assets between MNC entities. According to U.S. Treasury Regulations Section
1.482, Organization for Cooperation and Development (OECD) Transfer Pricing
Guidelines, and various other local country requirements, MNCs must document their
intra-company transactions in transfer pricing reports each year. Although
documentation requirements are country-specific, many countries follow the OECD
Guidelines for transfer pricing and most countries require that all material related-party
activities are documented contemporaneous with the firm’s tax filing.
MNCs hire consulting firms to document their intra-company transactions. The
consulting firms collect for the reports the MNC financials, organizational charts,
headcount by function, and intra-company agreements. The intra-company agreements
describe the transactional relationships, risks, economic owners of the intangible
property, payments, and contractual terms for the intra-firm transactions. The consultants
conduct interviews with senior managers and C-level executives to confirm whether the
activities were in accordance with the intra-company agreements. Any discrepancies are
documented in the reports. Since the reports document multi-year averages, MNCs
usually provide three years of consolidating income statement and balance sheet data.
The data for this research was put together under strict confidentiality. For this
reason, no company names or company-specific information can be identified. All
17
summary statistics show means, standard deviations, or other statistical measures, and
any descriptions are redacted or have the names changed to maintain anonymity.
Two datasets were compiled from the transfer pricing data: a MNC-level dataset
and a subsidiary-level dataset. The dataset is an unbalanced, longitudinal dataset since
the transfer pricing reports typically cover multiple years of data in order to take into
account business cycle and many of the MNCs repeatedly use the consulting firm for
services, making longer periods of time available. The sample is composed of 102
MNCs over the 1997-2012 time period on which the consulting firm had the most
comprehensive data. The sample contains a diverse group of MNCs, operating in a broad
range of industries and headquartered in a number of countries. The MNCs in the sample
had a combined total of 7,156 subsidiaries.
In addition to data on internal FSA ownership and contracting and licensing
arrangements, I collected patent data from the United States Patent and Trademark Office
(USPTO). I searched for all granted patents assigned to each MNC or to any subsidiary
within the MNC’s group. Following standard practice, the patents were matched to years
based on the filing date. I used the inventor city, state, and country information to match
each inventor location to a MNC subsidiary. Matching patents to inventor location
provides a much closer measure of subsidiary innovation than matching by assignees,
which can be biased by the intellectual property strategies of the MNCs. The MNCs in
the sample patented 50,934 patents over the sample period. 29,028 of the patents had
only one inventor and 10,711 patents had multiple inventors from the same subsidiary
location and therefore were coded to only one subsidiary. The remaining 11,195 patents
18
had inventors from more than one subsidiary location and were therefore coded to
multiple subsidiaries.
Coding of Data 3.1
Intangible Asset Definition. Transfer pricing reports discuss the MNC’s
“intangible assets” or “intellectual property.” The OECD defines an intangible asset as:
“not a physical asset or a financial asset, which is capable of being owned or
controlled for use in commercial activities, and whose use or transfer would be
compensated had it occurred in a transaction between independent parties in
comparable circumstances” (OECD 2013: 14).
Moreover, unique and valuable intangibles are:
“those intangibles (i) that are not comparable to intangibles used by or available to
parties to potentially comparable transactions, and (ii) whose use in business
operations (e.g. manufacturing, provision of services, marketing, sales or
administration) is expected to yield greater future economic benefits than would
be expected in the absence of the intangible” (OECD 2013).
OECD guidelines give several examples of the types of assets classified as
intangibles, including patents, know-how and trade secrets, trademarks and trade names
(OECD 2013). The definition of valuable intangible assets is closely aligned with the
strategy concept of core competencies and international business literature concept of
FSAs.
MNCs have different types of FSAs, ranging from product innovations to know-
how and formulas. The following table shows the different categories of intangible assets
listed in the transfer pricing reports for the MNCs in the sample.
19
Table 1: Coding of FSAs
FSA Category Example Description of FSA
Product [MNC's] intangible assets relate to its products. [MNC] has X product
families. [MNC] engages in developing and marketing products for
[industry]. [MNC]’s products include cutting edge [technology], are
highly integrated and offer high-quality performance. [MNC] expects that
its continued R&D efforts for [product line], which represents next-
generation technology based on the persisting legacy technology, will lead
to large sales growth.
Drug Discovery The profitability of [MNC] is determined by the successfulness of its drug
discovery, development, and commercialization capabilities. Drug
discovery activities include molecular discovery research.
Manufacturing Process [MNC] has proprietary manufacturing processes that provide it with
competitive advantages. Its manufacturing processes enable it to produce
high-quality products at prices lower than its competitors.
Procurement Know-How [MNC] Procurement Intangibles: [FSA owner] owns the rights to a
number of processes, procedures and proprietary tools used to optimize
the procurement of materials used in [MNC]’s operations. In order to
ensure high quality and consistency of suppliers, [FSA owner] has set
standards for all third parties that supply materials to affiliates located
worldwide, and developed a proprietary system for affiliates to evaluate,
select, monitor, and change suppliers.
Trademark and Marketing The brands owned by [MNC] are a significant source of value. [FSA
owner X] owns the worldwide rights to the valuable [Brand X] and
related trademarks.... [FSA owner Y] owns the rights to [Brand Y].
[Brand Y] is a highly regarded domestic and international brand in the X
industry. Third parties pay a considerable premium for the [Brand]
products due to the brand name.
Blends and Recipes [MNC]-owned intangible property is related to the blend formula and
recipes for creating the products sold by [MNC]. Primary processing
consists of blending chemicals in a batch run. Trade secrets and know-
how related to the blend formula and mixing procedures are also important
to the production process. The blends and formulas of [MNC]'s products
are proprietary to [FSA owner].
Retail Store Design Layout [MNC] Store Design Intangibles: [MNC] has specialized retail store
designs and layouts which [MNC] considers crucial for its brand image
and the ability of [MNC] to attract and retain customers.
20
Table 1: Coding of FSAs Continued.
FSA Category Example Description of FSA
Expertise and Know-How The services and know-how provided by the local subsidiaries drive
[MNC]'s profits. [MNC's] business requires highly skilled technical
experts. The local subsidiaries provide unique services, due to their
functional or geographic (depending on the entity) expertise. Because of
their unique skill and expertise, they are largely able to dictate the prices to
customers.
Customer Relationships Local customer relationships are important. [MNC's] trademark has no
value in a new market and therefore each local sales subsidiary must
create all customer relationships through its own intensive selling and
marketing activities. The sales subsidiaries make substantial investments in
relationship building and in the subsequent customer maintenance process.
The required time to build customer relationships varies by geographic
region, meaning that some sales subsidiaries incur losses for some time
while going through initial market penetration activities. The sales
subsidiaries have established loyal customer bases in their markets. These
relationships are valuable, locally developed intangible assets that are
owned by each local sales subsidiary.
Services [MNC]'s core strength is its service capabilities. [MNC] has the biggest
service network of any of the [industry] manufacturers. Customers in the
[X and Y industry sectors] seek [MNC]'s products due to its vast service
network, which provides [MNC] the ability to provide services anywhere
in the world.
Counter Example: MNC has
No Product Innovation FSA
[MNC] technology is relatively old and has not been updated or further
developed since [year]. Therefore, [MNC] currently incurs few to no
costs associated with developing, maintaining, or defending this
technology. Moreover, there is no market for the technology. The
products are considered standardized products.
21
Figure 2: Types of FSAs
Figure 2 provides a breakdown of the different types of FSAs owned by MNCs in
the sample. As shown in the pie chart, 42% of observations in the sample had product
innovation FSAs, 29% had trademark and marketing FSAs, 12% had manufacturing
processes, and the remaining had various other FSAs including distribution and retail,
drug discovery, and expertise and services, etc. MNCs sometimes indicate that they own
multiple FSAs. The most common overlap of FSA types was product innovation and
trademark intangibles.
Intangible Asset Owner Definition. I used the detailed information provided in the
transfer pricing reports to identify and code the intangible asset owners within the MNC.
The transfer pricing reports identify the intangible asset owners associated with each
intercompany transaction in order to determine which entity should receive any profits or
losses from the intra-company transactions. Intangible asset ownership under transfer
Drug Discovery4%
Product Innovation
42%
Manufacturing Processes
12%
Trademark & Marketing
29%
Customer Relationships
3%
Blends & Recipes
2%
Procurement Know how
1%Expertise &
Services5%
Distribution & Retail
2%
22
pricing guidelines is based on economic ownership. The intangible asset owners bear the
risk of intangible asset development or purchase previously developed intangible assets
from affiliates.
Intangible asset owners for each MNC were explicitly identified in the transfer
pricing reports. Ownership is also supported with the written legal contracts between
MNC entities and real flows within the firm. All transfer pricing reports for each MNC
were carefully reviewed and each intangible asset owner for each MNC was recorded, its
country of incorporation, and the functions performed by the intangible asset owner.
Tax haven Ownership. I code tax haven ownership as a binary indicator variable,
set equal to 0 if the country of incorporation of at least one of the MNC’s intangible asset
owners is a tax haven country. Tax haven countries were identified based on the OECD’s
list of tax havens (OECD, 2000). Some MNCs have FSA owners in tax haven countries
that perform substantial operational activities and significantly contribute to the
development of MNC FSAs. For this reason, I code two separate tax haven variables.
Operational Tax Haven is a binary indicator set equal to one if the FSA owner is located
in a tax haven country and performs R&D, distribution, manufacturing, or service
activities. Pure Tax Haven is a binary indicator set equal to one if the FSA owner is
located in a tax haven country and does not perform R&D, manufacturing, distribution,
sales and marketing activities, or financial trading (for banking firms).
FSA Ownership Structures 3.2
FSA Ownership Structures. MNC intangible asset ownership structures differ
along three dimensions: 1) the extent to which the ownership of their intangible assets is
centralized into one entity or dispersed across many entities, 2) whether ownership is
23
shared between entities or owned separately by entities, 3) whether the ownership is
located in tax havens.
From the information contained in the reports, I identified four types of intangible
asset ownership structures and coded each MNC according to its structure: 1) sole
ownership, where one entity owns all of the MNC’s intangible assets, 2) shared
ownership, where two or more entities share ownership of the intangible assets, and 3)
separate ownership where different intangible assets are separately owned by different
entities within the MNC, and 4) mixed ownership where some entities share ownership of
certain intangible assets and other entities own separate and distinct intangible assets.
Overlapping all of these types of structures is tax haven ownership, where one or more of
the intangible asset owners are located in tax haven locations.
24
Table 2: Coding of FSA Ownership Structures
Each structure has its advantages and disadvantages. The following provides an
overview of each structure.
3.2.1 Sole Ownership
Sole ownership occurs when one entity within the MNC owns the rights to all of
the MNC’s FSAs, regardless of where R&D, manufacturing, distribution, or marketing
Ownership Structure Definition Examples of Phrases Used in Reports
Sole* • • [Entity] is the owner of all intangible assets owned by [the
MNC].
• [Entity] owns, manages, and maintains [MNC's] portfolio
of intangible assets. These assets include, but are not
limited to, trademarks, process and information
technology, know-how, patents, industrial models, and all
other intellectual capital.
Shared • • [Entity] and [Entity] share ownership of the rights to all of
the techologies and trademarks associated with the
products owned by [MNC].
• Pursuant to the Cost Share Agreement, [Entity] and
[Entity] share all costs, risks, and rights to all of the
Company's intellectual property.
Separate • • [Entity A] owns the rights to [FSA 1]… [Entity B] owns
the rights to [FSA 2],…. and [Entity C] owns the rights to
[FSA 3].
• Each [distribution entities] own the rights to their local
market intangibles.
Mixed • • [Entity A] and [Entity B] share ownership of [X
FSAs]…[Entity C] owns the rights to [Y FSAs].
• [Entity A] is the economic owner and bears all costs and
risks of [MNC's] activities associated with [X FSAs].
Under the terms of a Cost Share Agreement, [Entity A]
and [Entity B] share the rights, risks and costs associated
with [FSAs]. [Entity D] owns the rights to [X FSAs].
All FSAs owned by the
MNC are owned by one
entity.
Two or more entities
within the firm share
ownership of the MNC’s
FSAs.
Two or more entities
within the MNC own the
rights to separate and
distinct MNC FSAs.
Two or more entities
share ownership of at
least one FSA and at
least one other entity
owns a separate and
distinct FSA.
* For MNCs that did not explicitly state that one entity owns all the FSAs, a MNC was also coded as having a
sole ownership structure if only one entity was named as the owner of FSAs in all of the MNC’s transfer pricing
reports.
25
activities occur within the firm. Figure 3 shows an example of a sole ownership
structure.
Figure 3: Sole Ownership Structure
In the above example, the parent is the only FSA owner of the MNC. The parent
owns all FSAs and contracts all other entities within the MNC to perform activities such
as R&D, manufacturing, and distribution. The FSA owner acts as a financial
intermediary in the trade relationships within the firm. Even though product flows
directly from the U.S. intermediate goods manufacturer to the U.K. finished goods
manufacturer, and from the U.K. finished goods manufacturer to the German distributor,
the FSA owner handles the financial flows by directly paying each entity for the services
performed and/or charging them for the goods received.
Sole ownership structures generate limited incentives within the firm since all but
one entity do not own FSAs (FSA users). FSA users earn a guaranteed income on their
activities, but do not have rights to residual profits earned from the FSA. Sole ownership
structures centralize operational risks such as development, market penetration,
manufacturing, and warranty risks into one entity. They also allow for centralized
Royalties for
FSAs
Risk & Residual
Income (Loss)
Specified Return
Product Flow
France U.S. South Korea
FSA User FSA User FSA User
R&D Manufacturing Manufacturing
Intermediate Goods Distribution
Product Flow Product Flow
Finished Goods
Parent
FSA Owner
Germany
Distribution
UK
Manufacturing
FSA User FSA User
26
coordination and control. Information flows from the FSA users to the one FSA owner
and from the one FSA owner to the FSA users. The information flows and connections
enable the FSA owner to direct the implementation of a new technology or innovation
across the group and to disseminate best practices learned from FSA users throughout the
firm. However, since all FSA users report to the one FSA owner, bounded rationality can
limit the FSA owner’s ability to identify the best investment opportunities. Therefore,
sole ownership should be particularly advantageous for small firms.
Sole ownership structures are administratively simple, easy to manage, and create
efficiency gains by having only one FSA-owning entity – usually the parent – contract
with the FSA users throughout the firm. Having a single FSA owner facilitates
monitoring the performance of FSA users and reduces the administrative complexity
associated with extensive intra-firm flows of FSA cross-licenses and royalties. Sole
ownership structures limit disputes and bargaining problems between MNC entities
regarding the allocation of returns since there is only one FSA owner within the firm.
The ease of management, low levels of bargaining problems, and ability to control and
leverage best practices make this structure also beneficial for very large firms that want to
reduce administrative complexity and have greater integration and control of their
operations.
3.2.2 Shared Ownership
Shared ownership of FSAs occurs when two or more entities co-own all of the
MNC’s FSAs. Under a shared ownership structure, the costs, risks, and benefits of the
FSA are shared between owners based on the relative contribution to the FSA,
27
geographical region, or field of use. Figure 4 provides an example of a shared ownership
structure based on geographical region.
Figure 4: Shared Ownership Structure
In the above example, the French entity owns the European rights to the FSAs and
the parent firm owns the U.S. and Asian rights to the FSAs. The FSA owners are
responsible for licensing the FSAs and contracting the FSA users for products or services
distributed within their region. Therefore, the French FSA owner contracts the U.S. and
U.K. entities to manufacture products and the German entity to distribute those products
in Europe. The U.S. contracts the South Korean entity to manufacture and distribute
product in Asia. The FSA owners share the costs of developing the FSAs and share the
risks and returns based on geographical regions. Thus, if the U.S. and Asian revenues
combined represent 65 percent of the total revenues of the MNC, the U.S. FSA owner
pays 65 percent of the development costs.
In shared ownership structures, the co-owners share incentives, control, and risks,
which can increase the MNC’s ability to expand geographically and engage in more risky
investment projects in the hopes of furthering growth. Shared ownership structures are
Royalties for FSAs
Risk & Residual
Income (Loss)
Specified Return
Product Flow
U.S.
FSA User
Manufacturing
Intermediate Goods
Product Flow Product Flow
Distribution
Responsible
Parent
FSA Owner
Germany
Distribution
South Korea
Manufacturing
for US & Asia
FSA User FSA User
Share FSA Development Costs
and Risks
France
FSA Owner
Responsible
Finished Goods
UK
Manufacturing
for Europe
FSA User
28
the used by the most R&D intensive firms. For these firms, joint ownership incentivizes
sharing knowledge and collaborating to increase total FSA value. Shared ownership
structures offer the greatest potential to coordinate within the firm in that different FSA
owners can manage regions, divisions or units, and work with other FSA owners to
decide the best strategic actions to take. Since different affiliates report to specific FSA
owners, information flows are more manageable and the problems of information
overload that can occur with only one FSA owner are mitigated.
An important disadvantage of shared ownership is that FSA co-owners share
power over decisions related to the FSA. Bargaining problems may arise between the co-
owners and differences of opinion can lead to non-optimal decision making. Shared
ownership structures are administratively more complex than sole ownership structures.
The co-owners contract different FSA users, so there is some duplication of efforts as
each FSA owner manages internal relationships.
3.2.3 Separate Ownership
Separate ownership occurs when two or more entities within the MNC own
different FSAs. For example, the Singapore affiliate of a consumer goods company may
own the rights to a technology and the UK affiliate may own the rights to a brand.
Alternatively, FSA ownership may be separated based on technologies, products,
localized relationships, brand names, or a mix of the above. The FSA owners license the
FSAs or contract the FSA users to perform activities. If the FSA owners want to use each
other’s FSAs, they cross license the rights to their FSAs. Figure 5 provides an example
of a separate ownership structure.
29
Figure 5: Separate Ownership Structure
In the above example of separate FSA ownership, there are four FSA owners.
Each FSA owner owns a distinct FSA. The parent firm owns the FSAs associated with
product group 1 and the French entity owns the FSAs associated with product group 2.
The U.S. entity owns the FSAs associated with its manufacturing processes. The U.S.
entity sells its intermediate goods directly to the U.K. manufacturing entity. The U.K.
manufacturing entity pays the U.S. intermediate manufacturer directly. Since the U.S.
manufacturing entity is an FSA owner, it is able to keep the profits associated with its
manufacturing activities. The South Korean entity owns the rights to the brand name and
marketing intangibles for the South Korean market. The South Korean entity sells
products from product groups 1 and 2. As a result, it pays both FSA owners royalties for
the products that it sells. Since the South Korean entity owns the rights to the brand
intangibles, it bears the risks associated with market penetration and brand development
in its region and is able to keep any profits from the brand intangibles. This incentivizes
the South Korean entity to create brand value in its region. The parent and French
entities are responsible for developing their own product groups and bear the risks as well
Royalties for FSAs
Risk & Residual
Income (Loss)
Specified Return
Product Flow
Product Flow
Manufacturing
Distribution
Product Groups
1 & 2
Product Flow
U.S.
1 & 2 1 & 2
FSA Owner
Manuf. Intangibles
Manufacturing
Intermediate Goods
Manufacturing
Product Groups Product Groups
FSA Owner
Product Group 1
Distribution
U.K. Germany
FSA User FSA User
France
FSA Owner
Product Group 2
Intangibles Intangibles
Parent
South Korea
FSA Owner
Brand Intangibles
30
as returns from their efforts. The U.K. manufacturer and German distributor do not own
the rights to any FSAs. Instead they earn set returns on their manufacturing costs.
By dispersing risk and control rights to FSA units throughout the firm, separate
ownership structures create market-like incentives. In comparison to sole ownership
structures, separate ownership structures can improve MNC investment decisions since
the entities that are best positioned to make decisions about the FSAs own and control
them. Separate ownership tends to create fewer bargaining problems than shared
ownership since each FSA owner has full control over an FSA and does not need another
party’s approval to make a decision.
Despite their advantages, separate ownership structures can increase the difficulty
of leveraging innovations and best practices across the firm. Information flows between
the FSA users and the particular FSA owner(s) with which they contract. If an FSA
owner were to use an innovation from another FSA owner, it would have to pay the other
FSA owner a royalty for the innovation. Therefore, separate FSA owners have the
incentive to develop innovations themselves and not to source knowledge from other
FSA owners. When information is shared, cross licenses need to be negotiated within the
firm. This can create bargaining problems and increase internal contracting costs.
Additionally, separate structures can make it difficult to inventory the kinds of
knowledge that reside with different FSA owners throughout the firm.
Separate ownership structures are more administratively complex than sole and
shared ownership structures since the different FSA owners act as “mini firms” within the
MNC. To the extent that there is operational overlap from having FSA users perform
activities for multiple FSA owners, it can become complex to track the internal
31
transactions and the royalties due to each FSA owner. The MNC may experience
reduced efficiency from duplication of efforts and administrative complexity.
3.2.4 Mixed Ownership
Mixed ownership occurs when two or more entities within the MNC share
ownership of at least one FSA and at least one entity within the MNC owns a separate
and distinct FSA. For example, in a consumer product company, one entity may own the
rights to old technology associated with the legacy product lines and two entities may
share the rights to new technology, which is associated with the new product lines.
Figure 6 provides an example of mixed ownership structure.
Figure 6: Mixed Ownership Structure
The mixed structure is basically a combination of both the shared and separate
structures. In the above example, the French entity and parent share ownership of the
MNC’s product intangibles based on geographical region. The U.S. manufacturer also
separately owns the manufacturing process intangibles and South Korea owns the brand
Royalties for FSAs
Risk &
Income (Loss)
Specified Return
Product Flow
U.S.
FSA Owner
Manuf. Intangibles
Manufacturing
Intermediate Goods
Product Flow Product Flow
South Korea
FSA Owner
Brand Intangibles
Manufacturing
Responsible for
Europe
Responsible for US
and Asia
Share FSA Development Costs and
Risks
FSA Owner FSA Owner
Product Intangibles Product Intangibles
France Parent
Finished Goods
Distribution
FSA User FSA User
Manufacturing Distribution
U.K. Germany
32
intangibles. Since the US manufacturer owns the manufacturing process, it directly sells
its manufactured goods to the UK manufacturer and keeps any profits from its activities.
The French FSA owner contracts the UK entity to manufacture product and the German
entity to distribute product within the European region. The parent licenses the product
intangibles to South Korea. South Korea, as an FSA owner bears the risks of market
penetration and brand development in its local market. It therefore keeps any residual
profits above the royalty paid to the parent for the product FSAs.
Mixed ownership structures provide the ability customize ownership of the FSAs
owned by the MNC. Mixed ownership structures provide market-like incentives for the
separate FSA owners within the firm and incentives to collaborate on innovation for the
FSA co-owners within the firm. Mixed ownership structures decentralize coordination as
different FSA owners and co-owners are responsible for different FSAs and activities.
Information flows between the FSA users and the particular FSA owner(s) with which
they contract.
Mixed ownership structures are the most operationally complex. Having different
FSA owners and co-owners means that the firm may experience problems with power
struggles and bargaining. Mixed ownership structures can have operational overlap if
multiple FSA owners contract with the same FSA users, which creates administrative
complexity for managing the different FSAs, activities, transactions, and returns across
the MNC. It is therefore not surprising that among the firms in the sample, very few
choose this type of structure to organize internal ownership rights to FSAs.
33
3.2.5 Tax Haven FSA Ownership
Tax haven FSA ownership occurs when MNCs locate all or part of their FSAs in
low tax jurisdictions. Examples of tax haven locations include Ireland, Luxembourg,
Hong Kong, Singapore, Bermuda, and Cayman Islands (OECD, 2000). Tax haven
countries typically implement a combination of policies designed to attract MNCs,
including low tax rates, minimal currency and banking controls, confidentiality, low
interest rates on loans, and high interest rates on deposits. There is variation in the extent
to which tax haven countries require that MNCs locate operations (“material substance”)
within their country. Some countries, such as Cayman Islands, Luxemburg, and
Barbados, have virtually no material substance requirements. Other countries, such as
Ireland, require that MNCs establish material substance by locating real operations within
the country.
Tax haven FSA ownership can occur in any of the FSA ownership structures. Not
all MNCs use tax havens to own FSAs. In the sample, 42 percent of MNCs have tax
haven FSA owners and 17 percent of the sample (38% of the tax haven FSA owners)
have pure tax haven FSA owners. When a tax haven FSA owner has operational
activities such as R&D, manufacturing, or distribution, the employees within the tax
haven entity may make strategic and operational decisions regarding the FSAs. When tax
haven FSA owners have no operational activities, the parent typically makes the strategic
decisions regarding the development and use of the FSAs. In the latter case, the primary
advantages of tax haven ownership are financial (reduced taxes or tax deferral) and/or
growth related.
The following figure contains a breakdown of the types of FSA ownership
34
structures used by MNCs, by industry.
Figure 7: FSA Ownership Structure by Industry
As shown in the figure above, the structures used by MNCs vary by industry.
MNCs in services industries and raw materials industries are more likely to use separate
structures than any other structures.
The MNCs that use the different types of structures vary greatly in characteristics.
The following table contains descriptive statistics on MNC characteristics, FSA
dimensions, and FSAs by FSA ownership structure. Superscripts denote significant
differences between structures. Each structure is numbered 1-4 (Sole, Shared, Separate,
and Mixed). Thus, the average R&D intensity for MNCs using a sole structure is 0.16.
The superscripts 2, 3, 4 indicate that the mean value of R&D intensity for sole ownership
is significantly different from the values for the three other types of structures. The R&D
intensity for MNCs using a Shared structure is 0.19 and the superscripts 3 and 4 indicate
this value is significantly different from the R&D intensity values of 0.08 and 0.10 for
Separate and Mixed ownership structures, respectively.
Agriculture
Mining & Construction
Food, Textile, Chem Mnf
Plastics, Metal, & Equip Mnf
Transportation & Comm Services
Wholesale & Retail Trade
Finance, Insurance & Real Estate
Business, Auto & Personal Services
Legal, Edu & Engineering Services
Sole
Shared
Separate
Mixed
0% 20% 40% 60% 80%
35
Table 3: Mean MNC Characteristics and FSAs by FSA Ownership Structure
All Mixed
MNC Characteristics (1) (2) (3) (4)
Revenue 13.92 12.622,3,4
13.413,4
14.704
15.83
Assets 14.43 13.082,3,4
13.813,4
15.314
16.31
Number of Subsidiaries 3.63 2.912,3,4
3.343,4
4.024
4.94
R&D Intensity 0.13 0.162,3,4
0.193,4
0.08 0.10
Profitability -0.05 -0.133,4
-0.234
0.06 0.12
Effective Tax Rate 0.26 0.25 0.24 0.28 0.19
Tax Haven FSA Owner 0.51 0.072,3,4
0.843,4
0.584
1.00
Pure Tax Haven FSA Owner 0.19 0.002,3,4
0.543
0.114
0.53
Age 41.66 32.462,3,4
24.923,4
57.564
19.34
Diversification 0.32 0.213,4
0.203,4
0.364
0.90
Number of M&As 1.14 0.473,4
0.633,4
1.68 2.06
US Country of Incorporation 0.75 0.853,4
0.813,4
0.68 0.59
EU Country of Incorporation 0.17 0.032,3,4
0.103,4
0.25 0.41
Other Countries of Incorporation 0.05 0.122,3,4
0.003,4
0.044
0.00
Contractibility
Tacit Scale 1.15 1.742,3
3.123,4
-0.144
1.43
Tacit 31.59 33.142,3,4
46.963,4
24.75 23.75
Codifiable 19.43 14.213,4
14.133,4
26.504
8.72
Independent Scale -0.68 -0.842,3,4
-1.533,4
-0.234
-0.48
Independent 6.02 4.733
6.744
6.804
3.78
Complementary 13.41 13.712,3,4
22.973,4
9.53 8.78
FSAs
Drug Discovery 0.08 0.023,4
0.033,4
0.104
0.28
Product Innovation 0.68 0.712,3,4
0.933
0.524
0.91
Manufacturing Processes 0.19 0.232
0.003,4
0.23 0.38
Procurement Know-how 0.02 0.022,4
0.003,4
0.044
0.00
Distribution 0.04 0.073,4
0.073,4
0.01 0.00
Trademark and Marketing 0.51 0.552,4
0.303,4
0.544
0.78
Blends and Recipes 0.04 0.003,4
0.003,4
0.084
0.00
Expertise and Services 0.09 0.032,3,4
0.003,4
0.174
0.00
Customer Relationships 0.05 0.032,3,4
0.003,4
0.094
0.00
Intermediate Product 0.45 0.572,3,4
0.773,4
0.24 0.38
Consumer Product 0.22 0.093,4
0.163,4
0.284
0.66
Software Programming 0.32 0.392,3
0.593,4
0.164
0.31
Sole Shared Separate
36
MNC Entity-Level Data 3.3
The MNC entities (parent and subsidiaries) in the sample come from a broad range
of countries. The following table shows the number of observations for each country and
the mean values of FSA ownership, entity size, R&D intensity, and role. The table shows
that FSA ownership is spread across a diverse set of countries.
37
Table 4: Subsidiary Descriptive Statistics by Country
Country Obs.
Tax
Haven
Country
FSA
Owner Parent Role
R&D
Intensity
Ln
Revenue
Ln
Assets
Argentina 148 0.09 0.00 1.22 0.00 5.27 5.10
Armenia 10 0.00 0.00 1.40 0.79 3.47 0.00
Australia 391 0.16 0.01 1.32 0.01 5.59 5.09
Austria 298 0.08 0.00 1.21 0.00 4.82 4.20
Bangladesh 4 0.00 0.00 1.00 0.00 5.69 5.27
Barbados 29 Yes 0.55 0.00 0.07 0.15 6.79 4.39
Belgium 535 0.03 0.00 1.26 0.01 5.37 5.48
Bermuda 52 Yes 0.37 0.00 0.17 0.00 7.44 3.06
Bolivia 1 0.00 0.00 1.00 0.00 0.34 0.00
Bosnia and Herzegovina 7 0.00 0.00 0.86 0.00 0.00 0.00
Brazil 231 0.04 0.00 1.22 0.01 4.96 4.44
Bulgaria 18 0.00 0.00 1.00 0.00 3.99 3.27
Canada 739 0.19 0.00 1.31 0.06 4.65 3.74
Cayman Islands 93 Yes 0.26 0.00 0.29 0.03 5.62 3.46
Channel Islands 12 Yes 0.00 0.00 0.00 0.00 -0.17 3.43
Chile 57 0.07 0.00 1.16 0.07 0.00 0.00
China 567 0.05 0.00 1.47 0.06 4.74 3.55
Colombia 87 0.07 0.00 1.23 0.00 5.28 4.93
Costa Rica 14 0.00 0.00 1.57 0.00 2.15 2.62
Croatia 62 0.00 0.00 0.98 0.03 3.75 4.13
Czech Republic 240 0.03 0.00 1.04 0.02 4.31 4.10
Denmark 239 0.13 0.03 1.27 0.03 4.31 4.28
Dominican Republic 7 Yes 0.00 0.00 1.00 0.00 1.22 0.00
Ecuador 12 0.00 0.00 1.00 0.00 0.18 2.50
Egypt 22 0.00 0.00 1.00 0.00 6.15 7.26
El Salvador 4 0.00 0.00 1.00 0.00 -1.25 0.00
Estonia 27 0.00 0.00 1.19 0.03 4.26 3.49
Ethiopia 2 0.00 0.00 1.00 0.00 5.73 0.00
Fiji 2 0.00 0.00 1.00 0.00 2.94 0.00
Finland 218 0.20 0.00 1.30 0.03 5.08 4.51
France 1,424 0.07 0.00 1.22 0.01 5.03 5.09
Germany 2,943 0.03 0.00 1.10 0.02 4.84 3.39
Ghana 7 0.00 0.00 1.57 0.00 5.09 5.64
Gibraltar 1 Yes 0.00 0.00 0.00 0.00 0.00 0.00
Greece 139 0.12 0.00 1.33 0.00 5.18 5.21
Guatemala 9 0.00 0.00 1.33 0.00 0.83 0.00
Honduras 7 0.00 0.00 1.00 0.00 1.06 3.52
Hong Kong 410 Yes 0.01 0.00 1.20 0.06 4.69 3.22
Hungary 194 0.03 0.00 1.30 0.00 3.88 3.81
Iceland 9 0.00 0.00 1.00 0.00 2.94 3.14
India 234 0.02 0.00 1.28 0.16 3.73 3.39
Indonesia 25 0.00 0.00 1.28 0.00 4.88 5.50
Ireland 799 Yes 0.12 0.01 0.83 0.03 4.70 4.91
Israel 85 0.15 0.02 1.54 0.17 3.85 3.19
Italy 651 0.10 0.00 1.29 0.00 5.16 4.73
Jamaica 10 0.00 0.00 1.00 0.00 3.92 6.55
Japan 365 0.05 0.01 1.47 0.04 5.51 3.97
38
Table 4: Subsidiaries by Country Continued
Country Obs.
Tax
Haven
Country
FSA
Owner Parent Role
R&D
Intensity
Ln
Revenue
Ln
Assets
Kenya 10 0.00 0.00 1.70 0.00 5.91 7.44
Latvia 36 0.00 0.00 1.00 0.00 4.31 4.92
Lithuania 16 0.00 0.00 1.00 0.00 4.85 4.17
Luxembourg 135 Yes 0.16 0.00 0.90 0.04 5.09 6.48
Malaysia 170 0.00 0.00 1.32 0.03 4.35 3.69
Malta 17 Yes 0.00 0.00 0.65 0.00 3.91 3.93
Mauritius 10 0.00 0.00 0.10 0.00 4.78 1.82
Mexico 363 0.12 0.00 1.15 0.00 4.14 3.02
Morocco 2 0.00 0.00 1.00 0.00 6.35 0.00
Netherlands 1,315 0.02 0.00 1.00 0.04 4.84 5.17
Netherlands Antilles 3 Yes 0.00 0.00 0.00 0.00 0.00 0.00
New Zealand 95 0.09 0.00 1.11 0.00 4.09 4.12
Nicaragua 2 0.00 0.00 1.50 0.05 0.00 0.00
Nigeria 9 0.00 0.00 1.00 0.00 5.54 5.43
Norway 181 0.08 0.00 1.25 0.00 4.71 4.07
Pakistan 27 0.00 0.00 1.26 0.00 5.99 6.19
Panama 11 Yes 0.00 0.00 1.00 0.00 2.39 3.88
Paraguay 1 0.00 0.00 1.00 0.01 0.02 0.00
Peru 40 0.00 0.00 1.23 0.00 3.53 5.37
Philippines 63 0.08 0.00 1.27 0.00 4.57 4.28
Poland 433 0.05 0.00 1.10 0.00 4.39 4.10
Portugal 164 0.09 0.00 1.26 0.01 4.70 4.34
Romania 165 0.13 0.00 1.03 0.01 2.29 2.68
Russian Federation 165 0.07 0.00 1.17 0.03 2.90 3.06
Saudi Arabia 2 0.00 0.00 2.00 0.00 0.00 0.75
Serbia * 27 0.00 0.00 0.96 0.00 1.73 2.40
Singapore 317 Yes 0.05 0.00 1.22 0.01 5.15 4.40
Slovakia 44 0.00 0.00 1.00 0.00 2.43 2.92
Slovenia 28 0.00 0.00 1.21 0.01 5.16 3.66
South Africa 99 0.04 0.00 1.20 0.00 3.75 2.50
South Korea 309 0.00 0.00 1.36 0.05 4.23 3.92
Spain 785 0.08 0.00 1.18 0.00 4.88 4.47
Sri Lanka 10 0.00 0.00 1.40 0.00 5.94 4.16
Sweden 670 0.07 0.01 1.13 0.02 4.48 4.67
Switzerland 451 Yes 0.24 0.04 1.07 0.13 5.22 4.05
Taiwan 222 0.02 0.00 1.38 0.10 3.53 2.87
Thailand 154 0.00 0.00 1.34 0.00 4.26 3.76
Trinidad & Tobago 15 0.00 0.00 1.47 0.01 3.43 3.81
Turkey 48 0.00 0.00 1.10 0.05 2.00 2.24
Uganda 2 0.00 0.00 1.00 0.00 6.13 5.82
Ukraine 61 0.03 0.00 1.11 0.00 3.41 3.60
United Arab Emirates 22 0.00 0.00 1.27 0.01 2.72 1.05
United Kingdom 8,781 0.06 0.00 1.05 0.02 4.97 3.95
United States of America 2,670 0.16 0.09 1.27 0.07 5.24 5.10
of America (Puerto Rico) 76 0.28 0.00 0.97 0.03 4.65 5.19
Uruguay 17 0.00 0.00 1.24 0.00 2.26 4.67
Venezuela 38 0.11 0.00 1.16 0.00 4.39 3.44
Vietnam 8 0.25 0.00 1.00 0.22 1.98 0.95
Virgin Islands (British) 6 Yes 0.00 0.00 0.00 0.00 -0.93 3.93
Zambia 6 0.00 0.00 1.00 0.00 3.66 3.06
Total 29,741 0.8 0.08 0.01 1.14 0.03 4.78 4.19
39
In the dataset, there were only 24 changes in FSA ownership (less than one
percent of observations). The following figure shows the number of changes by year.
Figure 8: Number of Changes in FSA Ownership by Year
The majority of changes in FSA ownership coincides with organizational
restructurings and occurs around the 2008 financial crisis. The number of changes
increases from four in 2006 and 2007 to eight in 2008, before dropping to two in 2009.
Approximately 24% of the changes were transferring ownership to operational non-tax
haven subsidiaries, another 24% transferred ownership to operational tax haven
subsidiaries, 36% were transferred to pure tax haven subsidiaries, and 16% were
transferred to the parent.
Table 5 contains the descriptive statistics on subsidiary, country, and MNC
characteristics by FSA users versus FSA owners.
0
1
2
3
4
5
6
7
8
9
2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010
Number of Changes
40
Table 5: Mean Value of Subsidiary Characteristics by FSA Ownership Type
The table contains the mean values of the various variables and the superscripts
indicate significant differences between FSA users and FSA owners. For example, the
mean number of patents produced by FSA users is .26 patents. The superscripts 2 and 3
indicate that .26 is significantly different from the patenting activity of the two FSA
owner types (shared FSA owner versus separate FSA owner). There are remarkable
differences between the FSA users and FSA owners. The FSA owners patent more
inventions, take on more roles, are larger and are much less likely to have been acquired
by the MNC. There are also large differences between FSA owners that share ownership
of FSAs versus those that do not (separate FSA owners). Shared FSA owners have a
All FSA User
Shared FSA
Owner
Separate
FSA Owner
Subsidiary (1) (2) (3)
Patents 0.47 0.262,3
5.003
2.93
Knowledge Sharing 0.20 0.142,3
2.533
0.74
Sole-Invented Patents 0.08 0.042,3
0.983
0.59
Pre-Sample Average Patents 0.09 0.072,3
0.823
0.35
R&D Intensity 0.03 0.032,3
0.203
0.05
Role 1.14 1.112,3
1.713
1.49
Subsidiary Size 4.76 4.612,3
7.883
5.91
M&A Dummy 0.26 0.262,3
0.073
0.23
Country
Country Effective Tax 0.22 0.222
0.203
0.22
Revealed Technological Advantage 0.85 0.843
0.793
0.93
Market Concentration 0.57 0.572
0.543
0.57
MNC
MNC Subsidiaries 182.90 189.002,3
68.173
102.19
MNC Ln Revenue 15.77 15.802,3
14.103
15.51
MNC Average Subsidiary Revenue 0.31 0.282,3
1.413
0.58
MNC Diversification 0.48 0.492,3
0.293
0.40
MNC R&D Intensity 0.07 0.072,3
0.183
0.06
41
much higher mean patenting, are more likely to engage in knowledge sharing than
separate FSA owners, have greater R&D intensity and subsidiary size. Interestingly,
shared FSA owners come from smaller MNCs than separate FSA owners.
Summary 3.4
There is large variation in MNC internal FSA ownership. Some MNCs have FSA
ownership centralized into one entity in the corporate group. Other MNCs have
ownership dispersed to many MNC entities. The internal organization of FSAs is not just
a question of centralization versus decentralization. MNCs choose whether to have MNC
entities separately own FSAs versus share FSAs. A surprisingly small number of MNCs
use mixed FSA ownership structures. The descriptive statistics show differences between
the MNCs that select different types of FSA ownership structures and differences
between FSA users and FSA owners. The purpose of this dissertation is to examine what
determines the different structures used by MNCs and explore the implications of FSA
ownership on subsidiary innovation
42
4. Inside the MNC: Structuring Ownership of Firm-Specific
Advantages
By
Catherine Magelssen
and
Susan Feinberg
ABSTRACT
This study examines how multinational corporations (MNCs) organize internal
ownership of their firm-specific advantages. Firm-specific advantages are the proprietary
assets that provide the firm with a competitive advantage. In contrast to the assumption
that firm-specific advantages are a public good within the firm, MNCs allocate economic
ownership of their firm-specific advantages to affiliates and/or the parent within the firm.
The MNC entities that own the firm-specific advantages internally contract or license
them to other MNC entities. This study identifies four different ways in which MNCs
choose to structure internal ownership of firm-specific advantages. Drawing on property
rights theory, it is argued that these structures are important in creating incentives and
facilitating coordination within the firm. The results suggest that MNCs with
independent and easily codifiable firm-specific advantages, such as trademarks, are more
likely to use ownership structures that provide market-like incentives. In contrast, MNCs
with knowledge-intensive, tacit firm-specific advantages are more likely to use
ownership structures that facilitate knowledge sharing and coordination within the firm.
43
Introduction 4.1
At the heart of the theory of the multinational firm is the idea that firms can exploit
a proprietary knowledge or resource more successfully by directly investing in a foreign
market than by licensing the asset to other firms. The inefficiencies associated with
licensing the asset often stem from a common set of problems related to contracting for
knowledge in conventional markets. Because markets for knowledge often fail, certain
transactions are better accomplished when they are “internalized” by firms (Caves, 1971;
Dunning, 1980; Rugman, 1981). Transaction cost economics and property rights theories
of the firm reach similar conclusions.
Economic theories of the firm have generally considered how contracting
difficulties affect whether transactions are conducted within or between firms. There is
widespread agreement in the fields of strategy and economics regarding the benefits of
“internalization” when contracts are difficult to write. However, once these transactions
are brought inside of firms, very little is known about how they are structured. Property
rights theory explicitly suggests that the same kinds of contracting difficulties that occur
at the boundaries of the firm are also likely to occur inside of firms (Grossman and Hart,
1986; Hart and Moore, 1990; Holmstrom and Roberts, 1998). Thus, the difficulties
inherent in contracting for knowledge in markets may not be resolved by internalization.
This paper examines how multinational corporations (MNCs) internally organize
ownership of their most important value-generating assets. These assets correspond to
what strategy and international business researchers variously call knowledge assets,
resources, competencies, and firm-specific advantages (FSAs). Although the theory of
the MNC suggests that firms establish foreign operations when their FSAs are not
44
suitable to contract or license in markets (Dunning, 1980), these same FSAs are
extensively contracted and licensed within MNCs. MNCs use internal contracts between
affiliates and/or the parent to assign economic ownership and control rights to FSAs
inside the firm.1 Since many MNCs grow by acquiring firms with existing FSAs or by
internally developing FSAs in foreign affiliates, firms face the organizational problem of
allocating FSA ownership and control rights across a diverse network of affiliates.
The need to create, maintain, and share FSAs within firms generates internal
motivation and coordination problems similar to those that exist at the boundaries of the
firm. To the extent that the FSAs involve tacit, incomplete, or shared knowledge and
specific investments are required, the same kinds of hold-up problems that occur in
market transactions are likely to occur within firms. This is particularly the case in
MNCs, where a firm’s network of foreign affiliates may be spread over many countries
with different customs, cultures, languages, and institutions. MNCs face additional
complications from the high costs of monitoring activities in a geographically dispersed
global network.
At the boundaries of the firm, the critical choice variable when contracts are hard
to write is whether to undertake a transaction within the firm, in markets, or through
some hybrid arrangement. Within firms, the critical choice variable with regard to
structuring internal transactions is whether to have all the FSAs owned by a single entity
and controlled centrally, to have multiple entities separately own the FSAs, or to have
multiple entities share ownership of the FSAs.
1 This analysis focuses on economic ownership, which is based on ultimately bearing the costs and risks of
the asset (Barzel, 1997).
45
Assigning economic ownership of FSAs to the parent and/or affiliate(s) reduces
free-rider problems that would arise if no entity within the firm had ownership rights to
the FSAs it created. If the FSAs were treated as pure public goods within the firm, there
would be no incentive to innovate, since the affiliates that bore no costs of developing the
FSAs would have access to the knowledge assets developed elsewhere within the firm.
When MNC entities own FSAs, they typically can determine the desired level of
investment in the FSAs and keep the profits earned from contracting or licensing the FSA
to other divisions of the firm. FSA owners are assigned responsibility and control over
decisions regarding the FSA and bear the operational risks associated with the
development, maintenance, and exploitation of the FSA.
Drawing on property rights theory, this study examines MNCs’ choices with
regard to FSA ownership structures. In general, it is expected that MNCs will select the
ownership structure that allocates FSA ownership and control rights to the division or
divisions whose marginal effort is most important with regard to developing,
maintaining, and reinvesting in the asset. For example, similar to Grossman and Hart’s
(1986) illustration of how client lists are owned in the insurance industry, FSAs, such as
customer relations and procurement know-how, are typically owned by the MNC unit
that deals directly with a particular set of customers or suppliers. However, in some
cases, the need for central coordination or the inability to monitor the actions of affiliates
will require greater involvement by the parent or headquarters in internal FSA
transactions. This study predicts that MNCs will choose to own FSAs that are more
“contractible” in the sense of being non-tacit, independent, and easy to describe, using
structures that engender more market-like incentives, but have less capacity for central
46
coordination. In contrast, MNCs will choose to own FSAs that are tacit or
complementary, and are therefore less “contractible,” using structures that allow for
greater coordination and control.
Using a confidential new panel data set on the internal economic ownership of
FSAs and transactions within MNCs, four different types of FSA ownership structures
used by MNCs are identified: 1) sole, 2) shared, 3) separate, and 4) mixed. First,
approximately 35% of MNCs in the sample choose “Sole” ownership structures in which
a single entity (usually the parent or headquarters) owns all of the FSAs of the MNC.
Second, 18% of MNCs choose “Shared” ownership structures in which two or more
entities co-own all of the FSAs. In shared structures, it is often the case that the parent or
headquarters shares ownership of all of the FSAs with an affiliate. A third structure
chosen by approximately 42% of MNCs is “Separate” ownership in which different
affiliates of the firm own different FSAs. Separate structures generate greater market-
like incentives within the firm than shared or sole structures, but can be difficult to
coordinate. Finally, about 5% of MNCs choose “Mixed” structures in which the
ownership of some FSAs is shared and other FSAs is separate. While mixed structures
are more customizable, they are also the most difficult to manage. As previously
discussed, different FSA ownership structures not only trade off incentives and control,
but they also necessitate very different patterns of knowledge and financial flows within
the firm. Once a particular structure is chosen, it is costly and difficult to change, leading
to a low incidence of switching observed in the data (less than 4% of observations).
The key contribution of this research is the ability to shed light on the structure of
transactions for knowledge within firms. Due to data limitations, previous empirical
47
research in economics and strategy has not been able to open up the black box of
internalized transactions. Although a great deal of research has examined the importance
of firm FSAs, little is known about the economic ownership of FSAs within the MNC.
This study identifies four ways in which MNCs structure ownership of their FSAs. The
four modes have different implications for control, coordination, incentives, and
knowledge sharing within the MNC. By examining how contracts for knowledge assets
are written and structured within firms, we gain insight into the ability of “incomplete
contracts” theories to explain internal firm behavior. Since contracts are so widely used
to delineate ownership of FSAs within firms, it seems likely that internalization is not
sufficient to deal with the difficulties inherent in contracting for knowledge.
From a policy standpoint, this research contributes to our understanding of the
kinds of benefits that foreign affiliates can bring to a local market. Affiliates that own the
MNC’s FSAs can accrue significant income from FSAs and have considerable strategic
determination over how the MNC’s key assets are developed, maintained, and deployed.
In this sense, ownership and control rights to FSAs are similar to subsidiary mandates
(e.g. Birkinshaw, 1996; Rugman, 1981). Additionally, some MNCs do choose ownership
structures as a way to minimize taxes. The results also shed light on the importance of
this aspect of MNC organization.
The next section describes FSAs in greater detail. The subsequent section presents
the literature review and predictions, followed by a description of the methods used to
test the predictions. The final section discusses the results and conclusions.
48
FSAs and FSA Ownership Structures 4.2
FSAs are the MNC’s proprietary assets that provide the firm with a competitive
advantage. These assets are unique company strengths and include a broad range
technological, manufacturing, marketing, human, and organizational competencies and
know how (Rugman, 1981; Rugman and Verbeke, 1992, 2001). FSAs may originate
from the parent or the network of affiliates of an MNC.
In order to gain and sustain a competitive advantage from the FSAs, firms need to
continuously develop and maintain their FSAs (Peteraf, 1993). This requires a careful
balance between the need to generate market-like incentives that encourage investment in
FSAs with the need to coordinate knowledge-generating activities and allocate capital to
the highest-return investments throughout the firm. To this end, MNCs use four different
internal FSA ownership structures.
Section 3.2 provides a detailed description of the four different FSA ownership
structures. As discussed in Section 3.2, each of the four types of structures used by
MNCs to organize ownership of their FSAs has different advantages and disadvantages.
The following table summarizes the important features of the four structures.
49
Table 6: Characteristics of FSA Ownership Structures
Sole Shared Separate Mixed
Control Over FSAs Centralized Shared De-centralized De-centralized
Risk Centralized into one
entity.
Shared by owners. De-centralized - risks
associated with
different activities
separately borne by
different owners.
De-centralized - mix of
shared and separately
held risks.
Incentives for Value
Creation
One entity has full
incentive.
Co-owners are
incentivized to
contribute to the value
creation process
together.
Owners are incentivised
to develop and create
their own FSAs.
Co-owners incentivised
to contribute together,
separate owners
incentivised to create
their own FSAs.
Financial Returns of
MNC
Centralized into one
entity.
Shared between a few
owners.
Dispersed to different
owners.
Dispersed to different
owners and co-owners.
Bargaining Problems Low - one entity owns
all FSAs.
High - hard to
distinguish contribution
of each co-owner and
shared control can lead
to conflict over future
direction of FSAs.
Medium - separate
owners within the MNC
hold power.
High - mix of co-
owners and separate
owners may try to use
power in bargaining.
Administrative
Complexity
Low - efficiency from
having one entity
contact with all entities.
Duplication of efforts
low.
Medium - co-owners
contract different
entities, with simplicity
of splitting any residual
returns. Duplication of
efforts medium.
High - complex to track
all contracting
relationships and that
each entity receives its
appropriate return.
Duplication of efforts
high.
High - complex to track
all contracting
relationships and that
each entity receives its
appropriate return.
Duplication of efforts
high.
Knowledge Flows Moderate Most Intensive Least Intensive Moderate
Information
Connections to FSA
Owners
High density -
connections between all
non-owners and one
owner.
Moderate density - non-
owner ties split amongst
the co-owners. Two-
way connections
between co-owners.
Low density - non-
owner ties to owners.
No/limited connections
between owners.
Moderate density - non-
owner ties split amongst
the owners and co-
owners. Two-way
connections between co-
owners, no/limited
connections between
separate owners.
Frictions in
Information/
Knowledge/Financial
Flows
Low - primary problem
bounded rationality.
Information overload of
owner and inability to
keep tabs on operations.
Lowest - by sharing
ownership, can delegate
responsibility for
portion of business to
co-owner. Reduces
information overload
problem. Co-owners
can then coordinate
business operations
together.
High - owners have
limited incentives to
share information or
resources with other
group members.
Moderate/High - While
co-owners have
incentives to share
information and
knowledge with other
co-owners, separate
owners do not have
incentives to share.
50
As discussed later in this paper, the regression models are at the enterprise level of
analysis, not at the affiliate level of analysis. Thus, it is beyond the scope of this
investigation to combine an analysis of MNC FSA ownership structures with location
choices. Therefore, the issue of locating ownership in a tax haven country is dealt with
by simultaneously estimating a model of structure choice with a model of the decision to
own FSAs in a tax haven. FSA characteristics are expected to be strongly associated with
FSA ownership structure choices. However, if the choice of ownership structure is
primarily motivated by the desire to minimize taxes, the simultaneous model that
includes the tax haven choice and the simple model of ownership structure choice should
have very different results.
Theoretical Background 4.3
To balance the need for motivation and coordination in FSA creation and
maintenance, firms use a number of mechanisms including job design, employee stock
ownership, and compensation (Holmstrom and Milgrom, 1991; Lambert, Larcker and
Weigelt, 1993; Wang et al, 2009). However, since performance can be difficult to
measure and monitoring is costly, Holmstrom and Milgrom (1994) argue that these
mechanisms enable only a very limited set of activities to be effectively rewarded. In
cases where these mechanisms fail to provide adequate incentives, Holmstrom and
Milgrom (1994) maintain that asset ownership can be a “broader, more powerful
incentive instrument” (p. 972) by enhancing the bargaining power of asset owners and
rewarding their investments in the asset. Property rights theory draws the same
51
conclusions, suggesting that asset ownership provides control over and income from the
asset and thus will be the most efficient way to incentivize investment.
As discussed in Chapter 3, FSA ownership structures vary in the degree of
centralized coordination they enable by allocating all FSA ownership rights to a single
unit versus the “market-like” incentives they create by assigning FSA ownership rights to
dispersed units within the firm. Because sole ownership structures are particularly
advantageous for small firms and the largest firms, the choice is likely to primarily be a
function of firm size, rather than FSA type. Similarly, since firms with mixed structures
choose to own their FSAs both separately and shared between units, this paper does not
have strong theoretical predictions at the firm level as to why mixed structures would be
chosen over shared or separate structures. Therefore the predictions focus on the choice
to own FSAs separately, or in structures that allow two or more MNC entities to share
FSA ownership.
4.3.1 Property Rights, FSA Contractibility, and Ownership Structure
Teece (1986) argues that firms exist to capture the returns from complementary
activities. However, when different units of the firm work together, conflicts and
bargaining problems can arise with regard to how to divide the returns – particularly
when the complementary activities involve knowledge creation. Holmstrom and Roberts
(1998) write, “Information and knowledge are at the heart of organizational design,
because they result in contractual and incentive problems that challenge both markets and
firms (p. 90).” Although integrated firms might be in a better position than independent
organizations at capturing the returns from knowledge transfers, “the problem of
knowledge transfers can be viewed as part of the more general problem of free-riding
52
when independent parties share a common asset” (Holmstrom and Roberts, 1998: p. 91).
When free-rider problems arise, incentives to invest in knowledge creation are muted.
Investment distortions arise, in part, because the ownership of non-human assets
affects the incentives to invest in human and other intangible assets such as marketing,
relationships, and knowledge creation (Hart and Moore, 1990; Feenstra and Hanson,
2005; Holmstrom and Roberts, 1998). It is costly and difficult to measure and evaluate
investments in intangibles, and frequently, it is not clear ex ante how much and what
kinds of investments in these assets are the most appropriate. Thus, contracts for the
provision of services related to human and intangible assets (FSAs) are nearly always
incomplete – even within firms (e.g., Grossman and Hart, 1986; Hart and Moore, 1990).
When it is not possible to write complete contracts and bargaining is costly, property
rights theory suggests that assigning FSA ownership to the relevant entities within the
firm will mitigate investment distortions. According to property rights theory, the
“relevant” entity to which FSA ownership and control rights should be assigned is the
entity that is in the best position to make important investments in the FSA (Grossman
and Hart, 1986).
Although assigning property rights to FSAs can help resolve hold-up problems
stemming from incomplete contracts, firms face several trade-offs related to using this
mechanism internally. First, assigning property rights can potentially impede knowledge
sharing within the firm to the extent that they serve as a “pay wall” to important firm
know-how. In this sense, assigning property rights within firms clearly diminishes some
of the potential benefits of internalization. For example, if a firm’s knowledge cannot
flow freely within its own boundaries internal spillovers from knowledge creation will
53
surely be attenuated. As discussed in the previous section, some FSA ownership
structures actually constrain the type and direction of knowledge sharing within firms.
Second, some kinds of knowledge assets are more “contractible” than others. When
complete contracts cannot be written that outline all possible rights and responsibilities of
FSA owners, the same kinds of contracting problems that occur at the boundaries of the
firm are likely to occur within firms.
The degree to which goods, services, and assets are “contractible” is related to how
easily they can be observed, separated, measured, or evaluated, and whether they require
relationship-specific investments and investments in intangibles. By definition, it is
easier to write complete contracts for goods, services, and assets that are more
“contractible.” Because markets provide stronger incentives for performance than firms,
transaction cost theory predicts that when complete contracts can be written, transactions
are more efficient when undertaken by independent parties in markets. In empirical
research on transaction cost theory, there is considerable evidence that contractible goods
or services are more likely to be exchanged in markets, rather than within firms (i.e.,
“bought” rather than “made”) (see, e.g. Anderson and Schmittlein, 1984; Masten, 1984;
LaFontaine and Slade, 2007). Property rights theory and empirical research comes to
analogous conclusions (see, e.g., Lerner and Malmendier, 2010; Elfenbein and Lerner,
2003; Kaplan and Stromberg, 2003). According to this literature, when complete
contracts cannot be written, market-like incentives can be achieved by assigning
“residual rights” ownership of important productive assets to the entity that is best
positioned to decide how the assets should be used. The assumption that problems
related to incomplete contracts can be resolved by exchange within firms is much
54
stronger in transaction cost theory than in property rights theory (Grossman and Hart,
1986).
Citing Kitch (1980), Burk and McDonnell (2007) suggest that the importance of
assigning property rights is particularly relevant for knowledge-intensive goods for which
complete contracts can almost never be written:
“[P]roperty rights in intangible information serve to solve the ‘disclosure
paradox’ recognized by Kenneth Arrow: once the innovator discloses its idea to
the would-be developer, there is no reason left for the latter to compensate the
former, but before disclosure of the idea, the developer does not know what it is
worth. This creates a potential stand-off, in which the innovator is unwilling to
disclose without assurances of payment, but the developer is unwilling to give
assurances until disclosure. This scenario can be identified as an extreme
variation on the hold-up problem in transaction cost economics... (p. 584).”
The difficulties in contracting for FSAs are similar to those that arise in
transactions for knowledge. Because FSAs are often intangible and are costly to develop
and maintain, firms have an interest in creating internal structures that provide high-
powered incentives to invest in FSAs. Separate FSA ownership structures share many of
the same features as markets. They provide strong incentives for FSA-owning units
within the MNC to create and maintain FSAs. These structures grant the FSA-owning
units decision and control rights to the FSAs and require these units to undertake the
associated risks. In return, FSA-owning units have the right to keep residual profits
associated with their endeavors. Because ownership of FSAs is dispersed throughout the
firm and control rights are disaggregated, separate structures trade off central control and
55
coordination for high-powered incentives to innovate. MNCs will be willing to make this
trade-off when FSAs are more contractible. When FSAs are easy to measure, observe,
and evaluate, structures that allocate FSA ownership and control rights to these entities
will generate fewer investment distortions.
When FSAs are difficult to measure, observe, or evaluate, or when FSAs require
knowledge inputs from more than one unit within the firm, structures that enable units to
collaborate and monitor each other’s inputs are likely to generate fewer investment
distortions. In contrast to separate structures, shared structures allow the MNC to reap
the maximum benefit from complementary assets. In this sense, shared structures trade
off market-like incentives for coordination and control. In shared structures, costs, risks,
and responsibilities for FSA creation and maintenance are shared by two or more units
within the firm and these units share residual profits generated by the FSAs. Knowledge
flows freely between units that co-create FSAs, and “pay walls” do not exist between
entities that share FSA ownership rights. Although property rights theory argues that
joint ownership of assets is never optimal in that it grants veto rights over assets to more
than one entity (Hart and Moore, 1990; Holmstrom and Roberts, 1998), it also it also
states that “assets that are worthless unless used together should never be separately
owned” (Holmstrom and Roberts, 1998: p.78).
Applying this insight to asset ownership within firms, firms are expected not to use
separate ownership structures when FSAs require inseparable contributions by more than
one entity within the firm. Contributions may be inseparable due to problems of
measurement and observability, or they may be inseparable in the sense of having no
stand-alone value. In cases of nonseparability, contracts detailing FSA ownership rights
56
will be nearly impossible to design and, therefore, separate FSA ownership structures will
distort incentives for investment.
Property rights theorists might argue that sole structures would be superior to
shared structures in cases of extreme non-contractibility. In markets, shared owners such
as alliance and joint venture partners can destroy a venture if conflicts arise with regard
to value creation and profit sharing. These conflicts are important reasons why so many
shared ownership arrangements like joint ventures ultimately fail (Park and Russo, 1996).
However, within firms, using sole FSA ownership structures when non-contractible
contributions are required by more than one division can be very costly with regard to the
lack of incentives these structures give to non-FSA-owning units. Hierarchies facilitate
the use of shared asset ownership in important ways that markets do not. First, the parent
entity can intervene to resolve intra-firm disputes, making it possible to avoid worst-case
scenarios that can occur when affiliates share ownership of intangibles. Second, MNC
parents are often FSA co-owners in shared structures, which further reduces the potential
for costly disputes (since one of the owners has the power of fiat). Indeed, recent
literature on property rights theory questions the premise that joint ownership is never
optimal and suggests that in some cases, the gains from sharing ownership outweigh the
potential costs (e.g., Hart 1995; Matouschek, 2004). Since hierarchies resolve many of
the potential hold-up costs related to sharing ownership, it is expected that firms will be
more likely to use shared structures if the firms’ FSAs are less contractible. Shared
ownership of complementary assets creates incentives for the FSA co-owners to invest to
increase the total value of the asset. Thus, Shared ownership structures are likely to be
superior to Sole ownership structures.
57
Hypotheses 4.4
Two dimensions of FSA contractibility— the independence versus
complementarity of the FSA and the codifiability versus tacitness of the FSA— are
expected to be associated with MNC decisions with regard to FSA ownership structures.
These are shown in the following matrix.
Table 7: FSA Contractibility
Structure Choice
FSA Contractibility Dimension Shared
Structure
Separate
Structure
Continuum 1:
Independent FSAs √
Versus
Complementary FSAs √
Continuum 2:
Tacit FSAs √
Versus
Codifiable FSAs √
4.4.1 Independence versus Complementarity
Independent assets are non-synergistic assets (Hart and Moore, 1990). They are
often idiosyncratic to different parties. When FSAs are “independent,” separate
ownership structures will be superior to all other ownership structures. Separate
ownership structures provide market-like incentives for developing and maintaining the
FSAs to the entities within the firm that are best positioned to invest in the FSAs.
Because no additional value is created from common ownership, the costs associated
with shared control are expected to outweigh the benefits.
58
In MNCs, there are two potentially different reasons why some FSAs are more
independent than others. First, some FSAs are clearly related to the local operations of
individual foreign affiliates and are not related with FSAs developed by other units of the
firm. This is often the case with FSAs like customer relationships and services. Second,
some FSAs do not require the contributions of more than one entity within the firm. This
is often the case with process-related FSAs that may be intentionally or accidentally
discovered as a result of variation in routines within a particular unit of the firm (Nelson
and Winter, 1982).
In contrast, complementary assets provide synergistic value. As discussed above,
when FSAs are complementary, separate ownership structures are never optimal. Thus,
the following hypothesis is advanced:
Hypothesis 1: MNCs with independent (complementary) FSAs will be
more likely to have Separate (Shared) ownership structures and less likely
to have Shared (Separate) ownership structures.
4.4.2 Tacitness versus Codifiability
Tacitness refers to the extent to which knowledge or some knowledge-based asset
is non-codifiable and requires application to be observed. Research on the knowledge-
based view of the firm (KBV) has given significant emphasis to the problems inherent in
transferring tacit knowledge (e.g., Grant, 1996; Kogut and Zander, 1992, 1993; Teece,
1977; Martin and Solomon, 2003). Tacitness is perhaps the most important feature of
knowledge contractibility. When knowledge is embedded in complex routines and
decision rules, the cost of transferring it can become prohibitively high, effectively
making it non-contractible. Tacit knowledge has been found to significantly increase the
59
costs of transferring knowledge abroad (Teece, 1977), to reduce the perceived ease of
internally transferring specialized marketing know-how (Simonin, 1999), and to increase
the time it takes to make new knowledge work well after it has been transferred
(Galbraith, 1990).
When the knowledge underlying FSAs is more tacit, FSAs are less contractible. In
such instances, MNCs will be less likely to use separate structures to allocate FSA
ownership rights. Since separate structures create mini firms within the MNC, these
structures enhance the difficulty of measuring and monitoring affiliates’ contributions
and the difficulty of transferring knowledge across the firm. To the extent that some
FSAs are more difficult to measure and evaluate in the first place, the problems inherent
in separate structures will be exacerbated.
Hypothesis 2: MNCs with tacit (codifiable) FSAs will be more likely to
have Shared (Separate) ownership structures and less likely to have
Separate (Shared) ownership structures.
Methods 4.5
4.5.1 Data
To examine the predicted relationships, a new confidential panel data set on
internal FSA ownership and transactions within MNCs was used. The dataset was
compiled from several sources. First, information on the types of FSAs owned by the
MNCs and the FSA ownership structures were taken from transfer pricing reports from a
consulting firm. MNCs are required by tax authorities to document their intra-firm
transactions each year in transfer pricing reports that determine whether or not the intra-
60
firm transactions are at market price.2 The transfer pricing reports discuss in detail the
contractual relationships and transactions between MNC entities, the FSAs owned by the
MNC entities, and the activities performed by each entity in the transactional
relationship. Second, MNC financial data was gathered from Bureau Van Dijk’s Orbis
database. The Orbis database contains income statement, balance sheet, business
segment, and industry information on public and private firms located worldwide. For
every observation, the Orbis data was verified with the same financial information
contained in the transfer pricing reports. In cases where data were missing or in the very
few cases in which the Orbis data and the transfer pricing data contained different values
for the same financial information, data from the transfer pricing reports was used.
Third, merger and acquisition data for each MNC was collected from Thomson Financial
Worldwide Merger and Acquisitions database. The database covers both public and
private acquisitions.
4.5.2 Sample
Before discussing the construction of the panel, it is worth noting that there are
obvious sample selection biases related to the data source. All MNCs in the sample
sought the services of a consulting firm that specializes in advising MNCs with regard to
2 According to U.S. Treasury Regulations Section 1.482, Organization for Cooperation and Development
(OECD) Transfer Pricing Guidelines and various other local country requirements, MNCs must document
each year their intra-firm transactions in transfer pricing reports. Although documentation requirements are
country-specific, many countries follow the OECD Guidelines for transfer pricing and most countries
require that all material related-party activities are documented contemporaneous with the firm’s tax filing
(E&Y, 2013). The purpose of transfer pricing reports is to test whether or not their intra-firm transactions
are at market price and then document the analysis and results in the reports. In order to test that the
internal transfer price for each transaction is consistent with the external market price, the activities
performed by the MNC entities, their risks, and the economic ownership of intangible assets must be taken
into account. Therefore, the transfer pricing reports provide detailed descriptions of the products and
services, intangible assets owned, and activities performed by the entities. Intra-firm agreements are
typically included as appendices to the reports.
61
transfer pricing strategies and other related activities. However, many MNCs seek the
services of such firms and all MNCs with material intra-firm transactions have to compile
transfer pricing reports. While many of the largest MNCs have well-staffed internal legal
and tax departments, they still seek the advice of consulting firms on transfer pricing
issues. Thus, the sample of firms is believed to not be unusual, given that they have all
chosen to be multinational and to have networks of foreign affiliates.
The raw sample contains data on 102 MNCs over the 2000-2012 time window.
Altogether there are 672 organization-year observations on the 102 MNCs. The number
of years of data for each MNC ranges from three to 13 years. Orbis had financial
information on 514 of the 672 MNC-year observations. In order to avoid unnecessary
loss of observations in the sample, data were entered from the company consolidated
financial statements used for the transfer pricing reports when Orbis data were missing.
Consolidated financial data were missing for 80 observations, leading to a total of 592
MNC-year observations after merging the financial and FSA ownership data. Lagging
the independent variables reduced the sample by 94 observations. The final sample
contains 498 MNC-year observations on 93 MNCs.
The data was coded and compiled under strict confidentiality. For this reason, no
company names or company-specific information can be identified. Therefore, only
means, standard deviations, percentages, and other statistical measures are reported and
all qualitative examples are redacted to disguise identity.
The sample comprises a diverse group of MNCs. The firms in the sample operate
in a broad range of industries including consumer goods, pharmaceutical, semiconductor,
retail, and financial service firms. Approximately 76 percent of the MNCs in the sample
62
were headquartered in the United States, 15 percent in Europe, two percent in Asia, and
the remaining seven percent were headquartered in other regions. A total of 20 percent of
the firms in the sample are private. There is large variation in the size of MNCs in the
sample, with a heavily right-tailed distribution. The sample average MNC revenue is
USD 11.0 billion, with a standard deviation of USD 37.9 billion.
4.5.3 Level of Analysis
The analyses are at the MNC-level, rather than the affiliate level. Thus, although
the data contains information on whether the parent and/or particular affiliates own FSAs
and the extent to which the FSAs are contracted out or licensed across the firm, all the
financial data, FSA data and MNC structure data are at the MNC-level.
4.5.4 Variables
Dependent Variable
FSA Ownership Structure is operationalized as a categorical variable, coded 1 for a
Sole ownership structure, 2 for a Shared ownership structure, 3 for a Separate ownership
structure, and 4 for a Mixed ownership structure. Four binary indicators for each
ownership type were also created. These mutually-exclusive categories were coded
based on the detailed descriptions of FSA ownership provided in the transfer pricing
reports. Chapter 3 contains the definitions used to code the ownership structures and
examples of the qualitative descriptions of ownership that MNCs include in their transfer
pricing reports.
Independent Variables
63
MNC FSAs and Contractibility Dimensions. The transfer pricing reports contain
lengthy qualitative descriptions of the FSAs owned by the MNC. Underpinning these
FSAs are flows of licensing revenues between MNC entities. Thus, transfer pricing
reports describe the bundle of skills, knowledge, routines, processes, technologies,
patents and other firm value-drivers that are owned or co-owned by one or more entities
and contracted or licensed to other entities within the firm.
There are several ways to create the key explanatory variables from the data used
in this research. First, one can identify the types of FSA(s) owned by each MNC using
the detailed descriptions. The problem with this approach – although this study reports
robustness results that use this approach – is that firms’ often identify more than one
FSA. For example, a firm might indicate that its primary source of value is its ability to
create new products as well as the brands and trademarks that it currently owns. In this
simple case, it is not clear whether the firm’s primary FSA is product innovation or the
brands and trademarks it owns or both. The initial version of this paper analyzed four
FSA categories (1) Manufacturing Processes, (2) Expertise and Relationships, (3) Product
Innovation, and (4) Brands. Chapter 3 Table 1 provides qualitative examples of MNCs’
FSA descriptions.3 Each MNC was assigned as many FSA categories as the descriptions
conveyed. Thus, the firm above would have been categorized as having “Product
Innovation” and “Brands” FSAs. This labeling scheme is problematic for two reasons.
First, in some cases there may be only one identifiable revenue flow (from licensing or
contracting) associated with the bundle of FSAs, thus naming more than one FSA
3 An alternative approach is assigning each firm to a unique FSA category based upon the description in the
transfer price reports. A drawback to this approach is that it is often not straightforward which skill, asset,
process, technology, etc. is the “primary” value-driver of the firm.
64
category causes identification problems. Second, from a theoretical standpoint, Product
Innovation was considered to be more tacit and Brands to be more codifiable, so it is not
clear which structure would be predicted from theory. Despite the problems with this
approach, it is used here primarily for the sake of illustrating interesting properties of the
data.
A second approach, used here, is to remain agnostic as to what the FSAs actually
are and instead use the FSA descriptions to code the features of the MNC’s bundle of
FSAs. Word counts were used to construct the two hypothesized contractibility
dimensions (tacit vs. codifiable and independent vs. complementary). A list of potential
words was created for each end of the two dimensions (tacit and codifiable for Tacit
Scale and independent and complementary for Independent Scale). The words and
phrases were selected to be consistent with prior research (e.g., Zander and Kogut, 1995;
Ambrosini and Bowman, 2001; etc.). Then the list of words was narrowed down through
several iterations of reading FSA descriptions and adding and removing words from the
lists. The lists were narrowed by eliminating duplicate words, words that were only used
once or twice, or words that were often used to mean many different things. Once the
word list was created, the relevant sections of the transfer pricing reports were searched
for each of the listed words in the descriptions to form preliminary counts of “Tacit,”
“Codifiable,” “Complementary,” and “Independent” words. For example, in the phrase,
“…these new innovations are fundamentally distinct from, and do not rely upon, the
technologies used in the past….” the word “innovations” would be counted as tacit and
the word “distinct” would be counted as independent. The count of words was then
cleaned by reviewing the text again to ensure that, for example, words like “suite”
65
referred to a suite of products or applications rather than an address. In cases where the
words were out of context, the count was changed to exclude the irrelevant observations.
The following table contains the final list of words.4
Table 8: Contractibility Dimension Word Counts
After finalizing the word counts, scales were created from the word counts for the
two contractibility dimensions (Jap, Robertson, and Hamilton, 2011). Independence
Scale, was calculated as the total count of independent words minus the total count of
4 In robustness tests, words that could be tied to an organizational structure were removed from the scale
such as cross-functional, different, business segment, and standalone. The results to the analyses did not
change by removing the words.
Tacit Total Count Codifiable Total Count
expertise 475 trademark 3021
experience 1039 trade name 704
know-how 340 logo 219
knowledge 441 blend 360
trade secret 115 formula 1778
explore 38 recipe 75
innovat 1089 compound 1379
technology 8717 manual 383
solutions 2906 patent 2257
complex 1584 schematic 120
Tacit Scale
Complementary Total Count Independent Total Count
collaborat 431 standalone 161
combin 2043 separate 905
integrat 2859 used only, used primarily in 35
common 497 distinct 220
cross-functional 78 specialized 473
bundle 64 custom 907
companion 29 differentiated 83
complement 349 diversified 85
unified 373 -specific 69
suite 383 business segment 320
Independent Scale
66
complementary words divided by the natural log of total words searched. A positive
value of Independence Scale indicates that the MNC’s bundle of FSAs is independent,
whereas a negative value indicates that the MNC’s FSAs are complementary. Similarly,
Tacit Scale is measured as the total count of tacit words minus the total count of
codifiable words divided by the natural log of total words searched. A positive value of
Tacit Scale indicates that the MNC’s FSAs are tacit, whereas a negative value indicates
that the MNC’s FSAs are codifiable.
Control Variables
Firm Size. Firm Size is operationalized as the natural log of the total number of
MNC subsidiaries, lagged by one year. Small MNCs are expected to be more likely to
have a Sole ownership structure and larger MNCs are expected to be more likely to have
Shared, Separate, or Mixed structures. The number of subsidiaries is used as a measure
of size rather than assets or sales because the complexity of the MNC rises with affiliate
network size. To show this, the squared number of MNC subsidiaries is also included.
As complexity increases with network size, the administrative costs of the more complex
structures rise dramatically. Therefore, the largest MNCs are expected to use sole
ownership structures to reduce organizational costs. Thus, the number of subsidiaries is
expected to be negatively associated with sole ownership, and the squared number of
subsidiaries is expected to be positively associated with sole ownership. Firm Size and
Size Squared are mean-centered in the regression estimates.
R&D Intensity. MNC R&D Intensity is included in order to take into account the
relationship between innovation activities and structure choice. R&D Intensity is
67
measured as total MNC R&D expenditures in the prior year divided by MNC total
revenue in the prior year. MNCs in the sample had an average R&D Intensity of 12
percent. This measure is expected to be positively associated with Shared and Mixed
ownership structures.
Diversification. A one-year lagged total entropy diversification measure is used to
capture the diversity of a firm’s activities (e.g. Bowen and Wiersema, 2005; Palepu
1985). The variable is calculated as follows:
Total Entropy = ∑
Si represents the MNC’s share of total sales in business segment i. N represents
the total number of business segments in which the MNC operates. The measure is
calculated using the business segment information from Orbis and, in the case where
Orbis data was unavailable, from the consolidated financial data in the transfer pricing
reports. This variable equals zero for single business firms and increases with greater
levels of diversification. Diversification is expected to be positively associated with
Separate ownership structures since diversified firms are more likely to have independent
FSAs.
Number of M&As. An acquisition event is defined in this study as an MNC
acquiring or merging 100 percent with a target firm.5 The variable is calculated as the
sum of the total number of acquisitions and/or mergers that an MNC made in the prior
year. On average, the MNCs in the sample engaged in one M&A in a given year, with a
standard deviation of 2.38. The number of acquisitions made by the MNC each year is
5 In results not reported herein, the alternative definition of acquiring greater than 5% of a company was
used. The alternative definition of acquisitions did not alter the results.
68
included as a control for reasons similar to diversification. Since post-merger integration
is costly and complex, MNCs that undertake a greater number of acquisitions are
expected to be more likely to allow acquired firms to manage their own FSAs. Thus, the
Number of M&As should be positively associated with Separate or Mixed structures.
Tax Haven Ownership. Tax Haven ownership is a binary indicator variable, set
equal to one if the MNC has at least one FSA owner incorporated in a tax haven. Tax
Haven countries were identified based on the list of tax haven countries (Dharmapala and
Hines, 2007). Additionally, the binary indicator Pure Tax Haven was created to indicate
whether the MNC had a tax haven FSA owner with no real operational activities.6
All regressions include controls for time using a time trend and industry using
dummy variables for Manufacturing, Service, and Other industries. All independent
variables are lagged by one year.
4.5.5 Estimation
The hypotheses focus on the degree to which MNCs are more likely to own
independent (versus complementary) FSAs in Separate or Shared structures and tacit
(versus codifiable) FSAs in Separate or Shared structures. Therefore the main analyses
focus on firms’ decisions to use Separate or Shared ownership structures.
The analysis has three parts. First, a probit model of the choice between Shared
and Separate structures is estimated, excluding firms that use Mixed or Sole structures.
6 The indicator for Operational Tax Haven is set equal to one if the tax haven FSA owner performed at
least one of the following functional activities: R&D, manufacturing, distribution, sales and marketing
activities, or financial trading (for banking firms). The binary indicator variable Pure Tax Haven is set
equal to one if the firm’s tax haven FSA owner did not physically perform operational activities, defined as
R&D, manufacturing, distribution, sales and marketing activities, or financial trading (for banking firms).
69
Second, using the same sub-sample, a bivariate probit model is estimated that combines
the choice between Shared versus Separate structures and the choice to have an FSA
owner located in a Tax Haven. It is expected that these are endogenous and interrelated
decisions. The bivariate probit, which is a variation of the standard Heckman model,
allows for estimating the two decisions together. It is expected that there will be different
results between the bivariate probit model that includes tax haven choice and the simple
probit model if the choice of ownership structure is primarily motivated by the desire to
minimize taxes.
Third, as an exploratory analysis, the multinomial logit is used to examine the
choice set of FSA ownership structures.7 All of the models have the following general
specification:
Yit = αi + β1Xi(t-1) + β2Zi(t-1) + φj + τt + εit
Yit is the discrete choice, either between Shared or Separate FSA ownership
structures or between all four structures in the multinomial logit estimates. The bivariate
probit model adds a second Yit which is set equal to 1 if the MNC has Tax Haven FSA
owner(s).
β1Xit is a matrix containing each MNC’s scores for the Independent Scale and the
Complementary Scale, and β2Zi(t-1) is a matrix containing lagged MNC characteristics, in
particular, Firm Size and Size Squared, R&D Intensity, Diversification, and the Number
7 The multinomial logit estimates are problematic due to the fact that (1) the choice processes for the
different structures are not the same (e.g., firms do not go directly from Sole to Mixed) (2) some FSA
ownership structures can be viewed as substitutes for each other (e.g., Mixed ownership structures might be
used as a substitute firms choosing to change from Shared (Separate) to Separate (Shared) ownership
structures). Thus, the inclusion of Mixed structures clearly violates IIA. Therefore this study reports
multinomial logit results on a truncated sub-sample of MNCs that excludes 37 firm-year observations of
Mixed ownership structures.
70
of M&As. φj and τt are industry and year controls, respectively. All errors are
clustered to account for repeat observations on MNCs.
Results 4.6
4.6.1 Descriptive Statistics
Table 9: Mean MNC Characteristics and FSAs by FSA Ownership Structure
The table above contains the mean values of MNC characteristics by FSA
ownership structure. Superscripts 1-4 denote significant differences between structures.
Thus, the statistics for revenue can be interpreted as follows. The average of the natural
log of revenue for MNCs using a sole structure is 12.62. The superscripts 2, 3, 4 indicate
that 12.62 is significantly different from the corresponding values for all of the other
structures. The average of the natural log of revenue for MNCs using a Shared structure
is 13.41 and the superscripts 3 and 4 indicate that 13.41 is significantly different from the
corresponding values in Separate and Mixed ownership structures. Looking down the
table, there are large differences in some of the variables for each structure. For example,
MNCs using Sole and Shared ownership structures are much more R&D intensive than
All Mixed
MNC Characteristics (1) (2) (3) (4)
Revenue 13.92 12.622,3,4
13.413,4
14.704
15.83
Number of Subsidiaries 3.63 2.912,3,4
3.343,4
4.024
4.94
R&D Intensity 0.13 0.162,3,4
0.193,4
0.08 0.10
Profitability -0.05 -0.133,4
-0.234
0.06 0.12
Effective Tax Rate 0.26 0.25 0.24 0.28 0.19
Diversification 0.32 0.213,4
0.203,4
0.364
0.90
Age 41.66 32.462,3,4
24.923,4
57.564
19.34
Sole Shared Separate
71
MNCs using Separate or Mixed structures. MNCs using Mixed ownership structures pay
the lowest average effective tax rate, and they are also the most profitable on average.
Table 10: Descriptive Statistics and Correlations8
Table 10 contains descriptive statistics and correlations for the full sample of
MNCs. The highest correlation is between Firm Size and the Number of M&As (r=0.44).
Therefore, M&As and diversification are entered into the regressions separately from
Firm Size.9
Chapter 3 – Figure 7 shows the distribution of FSA ownership structures by
industry. Interestingly, 100% of MNCs in the finance, insurance, and real estate
industries use Separate ownership structures. Similarly, approximately 50% of MNCs in
the wholesale trade and legal, education, and engineering services industries also use
Separate ownership structures. The very high prevalence of Separate ownership
structures in the services industries can be attributed to the fact that services are often
8 The correlation matrix is calculated using 2007 single-year data since standard cross-sectional correlations
assume independence across observations. The correlation matrices for the pooled sample and for the other
single years were consistent with those shown in the correlation table above. 9 Entering all variables into the equation does not change the results for the primary variables of interest.
Mean s.d. 1 2 3 4 5 6 7 8 9
1 Shared 0.21 0.41
2 Tacit Scale 1.19 3.48 0.25*
3 Independent Scale -0.70 1.24 -0.33* -0.35*
4 Firm Size 0.18 1.32 -0.15 -0.17 0.19
5 R&D Intensity 0.12 0.19 0.10 0.14 -0.21 -0.35*
6 Number of M&As 1.00 2.38 -0.08 0.02 0.10 0.44* -0.10
7 Diversification 0.29 0.45 -0.09 -0.01 0.08 0.28* -0.10 0.08
8 Manufacturing Dummy 0.66 0.48 -0.17 -0.05 0.00 0.09 0.09 -0.09 0.14
9 Other Industry Dummy 0.09 0.29 0.10 -0.14 0.08 -0.07 -0.20 -0.11 -0.10 -0.43*
10 Tax Haven Dummy 0.49 0.50 0.37* -0.07 -0.07 0.34* -0.22 0.18 -0.03 -0.11 0.00
* p<.05. Year=2007. Number of Observations 67.
72
independent and tied to local markets. Thus, Separate structures give FSA ownership
rights to the affiliates that are best positioned to invest in these competencies.
Figure 9: Ownership Structure by Type
Figure 9 presents the distribution of FSA ownership structures in the sample and
the proportion of firms using each structure that have a pure tax haven FSA owner. Not
surprisingly, MNCs with Sole structures almost never allocate FSA ownership rights to
pure tax haven affiliates. Pure tax haven use is much more prevalent in Shared
ownership structures (55% of observations) as compared to Separate ownership structures
(11% of observations). MNCs using Mixed ownership structures are the most prolific
users of pure tax haven subsidiaries (59% of observations).
0.0%
5.0%
10.0%
15.0%
20.0%
25.0%
30.0%
35.0%
40.0%
45.0%
Sole Shared Separate Mixed
Per
cen
tag
e
FSA Ownership Structure
FSA
Ownership
Structure
Portion of
MNCs with
Pure Tax
Haven
Owner(s)
73
Figure 10: Percentage of FSA Types Shared vs. Separate
Figure 10 shows the FSA categories and the degree to which they are owned in
Shared versus Separate structures. The bars on the chart represent the percentage of total
observations that have Shared versus Separate ownership structures. Some fascinating
results emerge. First, 100% of MNCs with Manufacturing Process or Expertise and
Relationship FSAs use Separate ownership structures! The FSAs with 100% Separate
ownership fit with Hypothesis 1 in that these FSAs are typically independent. Thus,
these FSAs are expected to be separately owned. Consistent with Hypothesis 2, MNCs
with tacit FSAs like product innovation are much more likely to use Shared than Separate
ownership structures.
As discussed above, three different estimation techniques were used to test the
hypotheses. Table 11 reports the probit results with the dependent variable set equal to 1
for Shared ownership structure. MNCs with Sole and Mixed ownership structures are
excluded from these estimates. Column 1 contains the control variables Firm Size, Size
Squared, R&D Intensity, the time trend and dummies for Manufacturing and Other
industries (Service industries is the referent category). Column 2 contains simple probit
estimates of the base model with the Tacit and Independent scales added. The pseudo-R2
more than doubles when the two FSA dimension scales are added. Column 3 shows the
Manufacturing Processes
Expertise and Relationships
Product Innovation
Trademark and Marketing
Shared Separate
100% 80% 60% 40% 20% 0 20% 40% 60% 80% 100%
74
estimates of the base model with Firm Size and Firm Size Squared removed and Number
of M&As and Diversification added. Columns 4 and 5 contain the bivariate probit
estimates with the same independent variables as Columns 2 and 3, respectively. Finally,
Columns 6 and 7 show the bivariate probit estimates using FSA categories as explanatory
variables rather than FSA dimensions. The last two regression models can only be
estimated using two of the four FSA categories since 100% of MNCs with
“Manufacturing Process” FSAs and “Expertise and Service” FSAs use Separate
structures.
75
4.6.2 Regression Results
Table 11: Probit and Bivariate Probit Results Predicting Shared FSA Ownership
The estimates in Table 11 provide strong support for Hypotheses 1 and 2. In the
probit estimates and bivariate probit estimates in Columns 2-5, MNCs with Independent
FSAs are more likely to use Separate, rather than Shared ownership structures (p<.01)
and MNCs with Tacit FSAs are more likely to use Shared rather than Separate ownership
structures (p<.05). The estimated coefficients and standard errors for the two
1 2 3 4 5 6 7
H1 (-) Independent Scale -0.66** -0.73** -0.71** -0.80**
(0.25) (0.25) (0.26) (0.26)
H2 (+) Tacit Scale 0.20* 0.22* 0.19* 0.21*
(0.09) (0.09) (0.09) (0.09)
H1 (-) Brand -0.80* -0.60
(0.39) (0.39)
H2 (+) Product 3.04*** 2.76***
(0.71) (0.61)
Firm Size -0.31* -0.17 -0.19 -0.28†
(0.14) (0.14) (0.14) (0.15)
Firm Size Squared -0.16** -0.12† -0.12* -0.16*
(0.06) (0.06) (0.06) (0.07)
R&D Intensity 0.27 -0.41 -0.38 -0.48 -0.44 -1.04† -0.68
(0.61) (0.50) (0.51) (0.50) (0.50) (0.60) (0.59)
M&As -0.16† -0.17† -0.14†
(0.09) (0.09) (0.08)
Diversification -0.36 -0.47 -0.52
(0.40) (0.47) (0.39)
Trend 0.07† 0.03 0.04 0.05 0.05 0.04 0.05
(0.04) (0.04) (0.04) (0.04) (0.04) (0.04) (0.04)
Manufacturing Industry Dummy -0.69 -0.81 -0.81 -0.90 -0.82 -0.49 -0.16
(0.43) (0.56) (0.53) (0.56) (0.55) (0.48) (0.51)
Other Industry Dummy -0.61 0.13 0.12 0.11 0.14 1.68* 1.49*
(0.64) (0.70) (0.68) (0.66) (0.63) (0.67) (0.65)
Constant -0.43 -0.79 -0.85 -0.84 -0.93 -2.58** -2.72**
(0.57) (0.63) (0.57) (0.64) (0.59) (0.93) (0.86)
Number of Observations 348 348 349 348 349 348 349
Pseudo R-Squared 0.15 0.37 0.37
Wald Chi-Squared 24*** 30*** 31*** 44*** 43*** 46*** 39***
Log Pseudolikelihood -184 -136 -136 -319 -324 -309 -327
Wald Test of rho=0 (Chi-Squared) 5.30* 5.81* 9.04** 6.59*
Robust, clustered standard errors are in parentheses. Two-tailed tests for variable coefficients.
†p<.10; * p<0.05; ** p<0.01; *** p<0.001
Probit Bivariate Probit
76
hypothesized scales show remarkable stability across the different probit and biprobit
regression estimates.
In general, smaller firms tend to have Shared rather than Separate ownership
structures, and R&D Intensity is not related to the choice of Shared and Separate FSA
ownership structures. MNCs that make a larger number of M&As are more likely to use
Separate ownership structures (p<.10, Columns 3 and 5), but Diversification is unrelated
to FSA ownership choice.
Turning to the estimates in Columns 6 and 7, MNCs with Product Innovation
FSAs are more likely to use Shared, rather than Separate ownership structures (p<.001)
and MNCs with Brand FSAs are somewhat less likely to use Shared ownership structures
(p<.05, Column 6; p>.10, Column 7). These results are consistent with the estimates in
Columns 2-5 in the sense that Product Innovation FSAs tend to be more tacit and less
independent, and Brand FSAs tend to be more codifiable and independent. However, the
estimates using FSA categories suffer from collinearity with the industry dummies, since
many of the FSAs tend to be prevalent in particular industries, so these results should be
interpreted with some caution.
The bivariate probit results in Columns 4 and 5 include a second set of regression
estimates for which the choice to have a tax haven subsidiary is the dependent variable.10
Although the estimates are not reported in the table, in all cases there was only one
significant association—a negative and significant relationship between the choice to
have a tax haven subsidiary and the Independent Scale (p<.05). This relationship is
10 The estimates in Columns 7 and 8 use the same bivariate probit regressors in the tax haven model. A
Likelihood Ratio Test for Independence of the two regressions was significant (p<.05), however the
bivariate probit specifications need further refinement.
77
somewhat complex. The tax haven FSA owner bears the costs and risks of developing
the complementary FSAs and earns revenues from licensing the FSAs throughout the
MNC. In some cases, these tax haven FSA owners are in countries like Ireland and
Switzerland in which subsidiaries perform real activities and have real strategic
responsibilities. In other cases, tax haven FSA co-owners are in pure tax havens like
Cayman Islands, and the MNC parent makes all strategic decisions. Perhaps
complementary FSAs provide greater opportunities for taking advantage of tax havens
since they can be leveraged across the MNC’s assets. The results suggest that the choice
to use Shared versus Separate structures and the choice to have a tax haven affiliate are
interrelated. However, the results also show that real considerations, such as the features
of an MNC’s FSAs as well as Firm Size and M&A activity are important correlates of
structure choice.
78
Table 12: Multinomial Logit Results
Table 12 reports the multinomial logit estimates. As discussed in footnote 7, these
estimates are exploratory and the sample is truncated by removing MNCs that use Mixed
FSA ownership structures. Despite this shortcoming, the multinomial logit results shed
light on interesting differences between MNCs using Sole ownership structures with only
one FSA-owning entity and MNCs using structures in which more than one entity owns
FSAs (Shared or Separate ownership structures). The results in Column 1 of Table 12 are
similar to the results in Table 11. Compared to MNCs that use Separate structures,
Shared
versus
Separate
Shared
versus
Sole
Separate
versus
Sole
1 2 3
H1 (-) Independent Scale -0.93† -0.12 0.80†
(0.50) (0.21) (0.47)
H2 (+) Tacit Scale 0.21* 0.11 -0.10
(0.10) (0.09) (0.08)
Firm Size -0.30 0.33 0.62**
(0.25) (0.26) (0.19)
Firm Size Squared -0.18 -0.36** -0.18
(0.12) (0.12) (0.12)
R&D Intensity -0.49 -0.55 -0.06
(0.68) (0.43) (0.42)
Trend 0.05 0.03 -0.02
(0.08) (0.09) (0.07)
Manufacturing Industry Dummy -0.71 -0.41 0.31
(0.85) (0.77) (0.53)
Other Industry Dummy 0.14 1.01 0.87
(1.10) (1.10) (0.88)
Constant -1.31 -0.12 1.19
(1.13) (1.08) (0.84)
Number of Observations 498
Wald Chi-Squared 56***
Pseudo R-Squared 0.21
Log Pseudolikelihood -412
†p<.10; * p<0.05; ** p<0.01; *** p<0.001
Robust, clustered standard errors are in parentheses. Two-tailed tests for
variable coefficients.
79
MNCs that choose Shared ownership structures have FSAs that are more Tacit (p<.05,
Column 1) and less Independent (p<.10, Column 1). In Column 2, the only significant
difference between MNCs that choose Shared versus Sole ownership structures is that the
latter tend to be used by the largest MNCs in the sample. This is consistent with the idea
that as firms grow to be extremely large, FSA ownership structures that allow for
multiple FSA owners with joint control rights and their respective contracting
arrangements simply become too complex and difficult to manage.
The estimates in Column 3 compare the choice of Sole and Separate FSA
ownership structures. Interestingly, MNCs that choose Separate structures are larger, on
average than MNCs that choose Sole structures (p<.01, Column 3). However, the very
largest MNCs are equally likely to choose both structure types. Finally, consistent with
property rights theory, MNCs with Independent FSAs are more likely to choose Separate,
rather than Sole FSA ownership structures (p<.10, Column 3). This suggests that when
FSAs have no synergistic value, firms choose disaggregated ownership structures that
provide market-like incentives for FSA development and maintenance to individual
affiliates throughout the MNC.
Several robustness tests were performed to check the sensitivity of the results.
First, additional controls were included in the regressions such as a binary indicator for
whether the firm is a public or a private firm, binary controls for whether the MNC is
incorporated in the US, Europe, or elsewhere, and a control for host country effective tax
rate. Second, alternative measures were applied for some of the controls in the
regressions. For example, MNC number of subsidiaries was replaced with MNC
revenues as well as MNC total assets, Diversification was replaced with the total number
80
of four-digit SIC codes in which the MNC operates and the Number of M&As was
replaced with M&A activity defined as acquiring greater than 5% of the company. Third,
highly correlated variables were entered into the regressions one at a time and also the
regressions were run without Firm Size Squared. The results of these additional analyses
on the theoretical variables of interest were consistent with the results reported herein.
Discussion 4.7
This study investigates how MNCs internally organize ownership of their FSAs.
The findings suggest that the problems associated with incomplete contracts are not fully
resolved by bringing an activity inside the firm. From the standpoint of property rights
and transaction cost theory, a key result of this research is that the same features of FSAs
that render them more or less contractible in markets also explain the internal ownership
structures chosen by firms.
In contrast to the assumption that knowledge assets are public goods within MNCs
(e.g., Ethier, 1986), firms establish internal economic ownership structures in which FSA
owners contract or license the FSAs to other affiliates and/or the parent. Such structures
allow firms to balance high-powered incentives that reward innovation with internal
coordination and control. This creates internal market-like transactions and structures
within the firm and, to some extent, limits the benefits of internalization, such as the
ability to freely access spillovers from internally created knowledge. Even within firms,
free rider problems disincentivize FSA creation and maintenance. Thus, some degree of
excludability is necessary for successful FSA development within firms. To this end,
MNCs choose different structures to delegate internal residual rights ownership of their
most important knowledge assets—including assets for which contracts are extremely
81
difficult to write. Although these assets are considered to be non-contractible in markets
(e.g., Dunning, 1980), they are extensively contracted and licensed within MNCs.
This study identifies four different types of FSA ownership structures: sole,
shared, separate, and mixed. It is argued that the reason that these FSA ownership
structures exist is to balance motivation and coordination problems associated with
contracting for knowledge. This research draws on property rights theory to understand
the decisions that MNCs make with regard to the internal ownership of FSAs. Consistent
with the predictions, MNCs with independent and easily codifiable FSAs are more likely
to use separate ownership structures that provide high-powered incentives. In contrast,
firms with complementary and tacit FSAs are more likely to use shared ownership
structures that facilitate knowledge flows and coordination within the firm.
When FSAs are highly contractible, MNCs use separate structures, or “mini firms
within the firm,” to organize internal FSA ownership. In separate structures, individual
units within the firm own and control the FSAs in which they are best positioned to
invest. Separate structures are the most common choice of all the MNCs in the sample
and have several interesting features. First, they are chosen by the vast majority of
MNCs in service sector industries. Second, compared to Shared ownership structures,
Separate structures are much less likely to be used by firms that make extensive use of
pure tax haven FSA owners. Since Separate structures involve the delegation of
ownership along with decision and control rights to MNC affiliates, it is perhaps not
surprising that these rights are not often granted to pure tax haven units that perform no
other functions for the MNC.
82
Third, Separate structures are chosen by 100% of MNCs with Manufacturing
Process and Expertise and Service FSAs. In addition to being independent, these FSAs
are likely to be organized in Separate structures because they tend to be localized. Future
drafts of this research will examine the association between the “local versus global”
dimension of FSAs and FSA ownership structure choices.
When FSAs are less contractible in the sense of being tacit or complementary,
MNCs choose Shared ownership structures. These structures are not very widely used by
firms in the sample – most likely due to the potential bargaining problems they create.
Shared structures allocate FSA ownership and control rights to two or more entities
within the firm, hypothetically giving all shared owners veto power over decisions related
to the FSA. The theoretical link between internal shared structures and the predictions of
property rights theory is complex. In general, property rights theory maintains that two
independent entities should not share ownership of complementary assets (Hart and
Moore, 1990).11
Rather, one entity or the other should own the assets and divide the
returns. Hart and Moore (1990) refer to this as “integration,” meaning only one entity
owns the asset. “Integration” is analogous to ownership in hierarchies rather than
markets. Within the firm, however, potential hold-up problems that arise due to the
shared veto issue are mitigated by the fact that the parent entity can always resolve
internal conflicts by fiat. Indeed, MNC parents are often FSA co-owners in the sample of
11 More recent work in property rights theory has taken issue with Hart and Moore’s position that joint asset
ownership is never optimal. Holmstrom and Roberts (1998) note that joint ownership may be more
efficient when investments by more than one entity improve non-human assets (p79). Within firms, this is
likely to be the case when investments are costly and capital intensive, requiring the participation of more
than one unit within the firm.
83
firms, giving them even more power to settle disputes. Thus, within firms, many of the
potential drawbacks of shared ownership in markets are no longer problematic.
A key finding in this research is that within firms, shared ownership structures are
much more likely to be used when FSAs are tacit and complementary. In markets, shared
structures have the potential to maximize conflicts, whereas within the firm, they allow
for close coordination, knowledge transfers, and shared incentives to innovate. In this
sense, the ability to manage joint ownership arrangements when FSAs are tacit or
complementary or in cases where contributions of individual units within the firm are
difficult to measure, is one of the most important advantages of organization in firms
rather than markets.
Finally, in exploratory results this study examines the correlates of Sole ownership
structures. Consistent with the idea that sole ownership structures can reduce complexity
for very large firms, the results indicate that Sole ownership structures are chosen by the
largest MNCs in the sample. Additionally, firms choose Separate structures rather than
Sole structures when FSAs are independent. In such instances, Separate structures
provide better disaggregated incentives to develop and maintain the MNC’s FSAs.
It is important to note that all FSA ownership structures are the result of a two-step
process. First, ownership is “integrated” in the sense of being internalized within firms.
Second, the choice is made within the firm to integrate FSA ownership (Sole ownership),
share ownership between various units (Shared ownership), distribute FSA ownership to
“mini firms within the firm" (Separate ownership) or use some hybrid of the latter two
structures (Mixed ownership).
84
The results indicate that the choice of Shared versus Separate ownership
structures is interrelated with the choice to have tax haven subsidiaries. The majority of
MNCs in the sample have tax haven subsidiaries, and many have tax haven FSA owners.
MNCs with Shared ownership structures have a larger proportion of pure tax haven FSA
owners than MNCs with any of the other three structures. It therefore is no surprise that
the Internal Revenue Service classifies shared ownership between US and foreign MNC
entities as a Tier 1 tax issue. The finding that MNCs with complementary FSAs are more
likely to have tax haven FSA owners suggests that FSA characteristics may be associated
with a firm’s ability to take advantage of tax havens.
Although a detailed analysis of firm strategy and its relation to FSA ownership
structure choice is beyond the scope of this paper, the results seem to suggest that Shared
structures are used by relatively young, innovative MNCs in high growth industries.
Compared to MNCs with Sole and Separate structures, MNCs that use Shared structures
are significantly younger and less profitable. Compared only to MNCs that use Separate
structures, MNCs that use Shared structures have fewer subsidiaries and undertake fewer
M&As. MNCs with shared ownership structures are much more prevalent among MNCs
with Product Innovation FSAs in areas such as software programming. These stylized
facts, together with the much greater use of pure tax haven FSA owners suggests that
MNCs with Shared structures may channel the revenues from FSA ownership to tax
havens in order to fuel global growth.
FSA ownership has implications for policymakers and managers alike. For
policymakers, internal FSA ownership and contract and licensing arrangements have
been the subject of much scrutiny lately by legislatures and governments around the
85
globe (e.g. Levin and McCain, 2013; Bergin, 2012; Thompson, 2012). FSA ownership
has a significant effect on government revenues. Understanding the purpose and the
factors that drive the selection of FSA ownership structures can provide insight into the
types of policies that can attract MNCs to locate FSA ownership within a country. While
firms vary in the extent to which they locate FSA ownership in tax havens, it is important
to understand the operational antecedents of this choice.
From a managerial standpoint, the results indicate a clear association between the
characteristics of FSAs owned by the firm and the internal ownership structures used to
manage them. The different types of FSA ownership structures create different linkages
across the units of the firm in terms of knowledge flows and financial flows. The
creation and maintenance of tacit, knowledge-intensive FSAs require diverse insights and
collaboration. The finding that MNCs use shared ownership structures when their FSAs
are tacit is consistent with the idea that shared ownership structures provide better
mechanisms for coordination and knowledge transfer and are feasible to implement
within firms.
As previously mentioned, future versions of this research will incorporate several
refinements such as the “local versus global” dimension of FSAs. The present draft of
this research is limited by the fact that Sole ownership structures were only examined in
exploratory analyses. Sole structures are expected to be predictable by the interactions of
different FSA characteristics. For example, Sole ownership structures are expected to be
chosen when FSAs are codifiable and global. In such instances, Sole structures allow the
MNC parent to more tightly coordinate activities associated with codifiable but global
FSAs and transfer knowledge to diverse units within the firm. Testing this theory would
86
require creating an additional “local versus global” FSA dimension. Such an approach
will be used in the refinement of this research.
A second important future refinement is to better estimate a joint model of FSA
ownership structure and the choice to use a tax haven FSA owner. The results suggest
that these two decisions are interrelated, but a much stronger set of predictors is needed
for tax haven FSA ownership. The endogeneity of these decisions creates complex
problems with regard to pinning down the underlying structural relationships between tax
haven use and ownership structure choice.
In addition to refining the current preliminary draft of this research, there are a
large number of related research questions to investigate, which should provide further
insight into MNCs’ FSA ownership structure choices. First, future research will examine
firms’ decisions to switch from one ownership structure to another. Although this is a
relatively rare event, interesting patterns appear in the switching data such as the fact that
MNCs with Shared ownership structures never switch to Separate ownership structures,
but the reverse does not apply. An investigation of structure changes is beyond the scope
of the present research.
Second, future research at the affiliate level will study the relationships between
ownership structures and flows of goods and services within the firm. Previous research
suggests that intra-firm trade is knowledge-intensive and complex to organize within
firms (Feinberg and Keane, 2006); however detailed data on transactions within MNCs
have not been previously available to researchers. Third, future research will investigate
MNC structure choices at the affiliate level. A more disaggregated investigation of FSA
ownership structures will allow for examination of the degree to which these are
87
associated with location choices. Such an analysis would contribute to both the product
mandate literature and to the location choice literature. This unique data on ownership
and related transactions within MNCs can shed light on basic questions related to the
theory and management of multinational firms.
88
5. Subsidiary Ownership of FSAs and Innovation
ABSTRACT
This study examines the relationship between FSA ownership and subsidiary
innovation. Innovation is inherently difficult to monitor and control. Property rights
theory suggests that ownership provides two incentives for investing in the creation of the
asset: 1) the ability to appropriate income from the innovations created, and 2) the ability
to control both the asset and its future direction of development. The results suggest that
subsidiary FSA ownership is positively associated with innovation, and that transferring
ownership away from the subsidiaries that create the FSAs has a negative effect on
innovation. I also explore the effects of having tax haven subsidiaries own the rights to
MNC FSAs and find that subsidiaries that perform activities for pure tax havens innovate
less than those contracted by the parent.
Introduction 5.1
It is estimated that the U.S. government loses over $60 billion a year in tax revenue
due to corporate tax evasion (Bloomberg, 2010). Tim Cook, the CEO of Apple,
executives from Caterpillar, Inc., Starbucks, Amazon, etc., have all been called to testify
before government bodies in the United States and abroad, regarding their companies’
use of tax havens. As noted by a Wall Street Journal article, one key mechanism through
which firms are able to avoid taxes is by transferring economic ownership of their
intellectual property to subsidiaries located in tax havens (Wall Street Journal, 2002).
For example, Starbucks’ European subsidiaries license the rights to use Starbucks brands,
products, and processes from their Netherlands subsidiary, Starbucks Coffee EMEA BV
89
(Reuters, 2012). Because the Starbucks Netherlands subsidiary owns the economic rights
to Starbuck’s intellectual property, it is legally entitled to the profits from the intellectual
property. Management theorists have generally assumed that a firm’s key value
generating assets, or firm-specific advantages (FSAs) are public goods within the firm
(e.g. Ethier, 1986; Kogut and Zander, 1992). However, multinational firms (MNCs)
allocate ownership of their FSAs to subsidiaries and or the parent, who bear the risk of
developing the FSAs, maintain control rights to the FSAs, extensively contract and
license the FSAs to other entities within the firm, and receive the income or losses from
the FSAs. Subsidiary ownership of FSAs not only affects where tax payments and profits
reside within the MNC, it affects subsidiary-level control over assets and resources, and
incentives for the maintenance and creation of future FSAs. The allocation of FSA
ownership within the MNC can have important consequences for MNC innovation,
growth, and profitability. Despite its importance, relatively little is known about the
effects of the economic ownership of FSAs within the MNC. No doubt the absence of
research on this topic is mainly due to the lack of available data.
Using a new, confidential dataset on 102 MNCs and their subsidiaries, I examine
the ramifications of internal FSA ownership on subsidiary technological innovation. I
explore whether firms like Starbucks that strip FSA ownership rights from value
generating subsidiaries and give those rights to tax haven subsidiaries within the firm
creates incentive problems with regard to future innovative activity.
Subsidiaries play a key role in innovation, and it is generally to the benefit of the
MNC to allocate FSA ownership to the entity creating the FSA. Researchers have noted
the importance of foreign subsidiaries in the creation of FSAs for MNCs (e.g. Cantwell
90
1995; Dunning and Narula 1995; Kuemmerle, 1997; von Zedtwitz and Gassmann, 2002).
Subsidiaries are able to draw on their diverse local knowledge, skills, and expertise to
create innovations. Subsidiaries are not just locations to tap into the local expertise, but
are hubs for value creation within the firm (Birkinshaw et al, 1998).
The ability to share and leverage FSAs across complimentary activities and
geographically dispersed locations is an advantage of the firm. With the advantage,
however, comes the organizational problem of allocating control over the assets and
structuring incentives for the ongoing creation and maintenance of the FSAs. Since the
subsidiaries that create the FSAs may be different from those that use them, MNCs face
the organizational problem of allocating the ownership rights and control to the various
profit-generating FSAs within the firm.
In an exchange relationship, the party whose activity makes the largest
contribution to the creation and maintenance of the asset should own the asset (Grossman
and Hart, 1986; Hart and Moore, 1990). Transferring FSA ownership away from value
generating subsidiaries means that the MNC is also transferring the risk and rewards of
development away from the centers of innovation within the MNC. Chapter 4 suggested
that MNCs form different internal market structures to manage the way knowledge is
created and shared within the firm. Firms face tradeoffs in choosing both the FSA
structures and FSA ownership. While reducing tax expenditures is one reason for
assigning subsidiary FSA ownership, operational factors, such as innovation activities,
subsidiary role, MNC growth, and administrative complexity are also important in
determining FSA ownership. Transferring FSA ownership within the firm is costly and is
heavily scrutinized by tax authorities. The structuring of FSA ownership involves large,
91
often one-time decisions about a subsidiary owning FSAs. Because the FSA ownership
structures are long lasting and there are conflicting priorities that MNCs must balance,
some MNCs may have sub-optimal structures in place or structures that have unintended
consequences.
The purpose of this paper is to examine the effects of subsidiary-level FSA
ownership on subsidiary technological innovation. Ownership provides two incentives
for investing in the maintenance and creation of FSAs: 1) the ability to appropriate
income from the innovations created, and 2) the ability to control both the asset and its
future direction of development. Subsidiary FSA ownership is predicted to be positively
associated with innovation. As the vast majority of subsidiaries within the firm are FSA
users that are contracted by FSA owners to perform various activities, I further examine
the contracting relationships within the firm. Subsidiaries that are contracted by the
parent have increased incentives for advancement within the MNC and organizational
identification, which will motivate them to innovate. Thus, subsidiaries that contract
with the parent are predicted to be positively associated with innovation. In contrast,
subsidiaries that are contracted by pure tax haven FSA owners may become disenchanted
by having to give the fruits of their labor to another subsidiary whose activities do not
directly contribute to the development of the FSAs. Therefore, it is expected that
subsidiaries that are contracted by pure tax haven FSA owners will be associated with
reduced innovation.
To test the predictions, I compiled a subsidiary-level dataset on the economic
ownership of FSAs and subsidiary contracting and licensing relationships of 95 MNCs in
an unbalanced, panel dataset from 1997 to 2011. I combined the data with data from
92
Bureau Van Dijk’s Orbis database and patent data from United States Patent Trademark
Office (USPTO). The dataset contains information on 7,156 subsidiaries and their
50,934 patents over the 1997-2011 time period. The results to the analyses suggest that
FSA ownership is important for subsidiary innovation. Subsidiaries that own the rights to
the FSAs they create are significantly more likely to produce technological innovations.
Moreover, transferring ownership away from a subsidiary significantly reduces its
innovation. While contracting with a parent is positively associated with innovation,
contracting with a pure tax haven entity has a negative effect on subsidiary innovation.
This paper makes several important contributions to the literature on the MNC and
subsidiary innovation. First, it contributes to the theory of the multinational firm by
showing how the internal governance of FSAs affects the creation of future FSAs. The
free-flow of knowledge within the firm has been held as a key advantage of the firm
(Ethier, 1986; Hymer, 1960; Caves, 1971). However, within the MNC, knowledge is
bought and sold amongst the subsidiaries and/or parent. Assigning FSA ownership
within the MNC creates internal markets for knowledge, which can enhance innovation
within the firm. Second, this research contributes to the subsidiary entrepreneurship
literature by introducing economic ownership of FSAs as a means of establishing
entrepreneurial subsidiaries within the MNC. The FSA owners bear the risk and rewards
of not just their own operations, but also of other subsidiary operations. Third, I
contribute to the literature on subsidiary innovation by examining the effects of FSA
ownership on subsidiary innovation. Although FSAs have long held a central role in the
theory of the MNC, the effects of FSA ownership on subsidiary innovation remains
unexamined. Research on subsidiary innovation focuses primarily on subsidiary
93
capabilities and knowledge transfer and diffusion (e.g. Frost, Birkinshaw and Ensign,
2002; Almeida and Phene, 2004). Fourth, I go beyond transfer pricing research’s focus
on profit shifting and tax avoidance (e.g. Grubert, 2003; Mutti and Grubert, 2008;
Dischinger and Riedel, 2010) by showing how tax haven subsidiaries and FSA ownership
can affect real operations.
Contextual Background 5.2
Within the MNC, the parent and/or subsidiaries (entities) that own the FSAs (FSA
owners) internally contract other entities (FSA users) within the firm to perform activities
such as research and development, manufacturing, and distribution and pay the FSA users
a guaranteed return for their activities. MNCs use legal, written contracts between the
subsidiaries and/or parent to specify and solidify their intra-firm relationships. The
internal contracts outline the activities to be performed, the rights, risks borne, and
payment terms of the entities. Some MNCs have FSA owners that perform operational
activities, whereas others have “pure tax havens” – subsidiaries that are located in tax
haven countries, own the economic rights to FSAs, and do not perform operational
activities. The FSA owners typically maintain rights control over the use and
development of the FSAs. In the case of pure tax haven FSA owners, which tend to have
minimal employees, the parent typically controls the FSAs while the tax haven subsidiary
keeps the profits.
Before proceeding, it is useful to provide a brief example of the internal
contracting and licensing arrangements. ABC Transmissions is a transmission company
headquartered in the U.S. Its German subsidiary is the economic owner of the
transmission technology. The Germany subsidiary contracts the UK subsidiary to
94
develop transmissions for small vehicles and promises to pay the UK subsidiary a 7%
mark-up on its R&D expenses. If the UK subsidiary fails to create a new product or takes
an extra two years to do so, the U.K. subsidiary still receives the 7% return on its R&D
expenses and the German subsidiary incurs the loss associated with the R&D activities.
If the UK subsidiary creates the latest and greatest new transmission, the U.K. subsidiary
continues to receive the 7% mark-up on its R&D costs. The German subsidiary, as the
economic owner, receives any income from the new transmission above what it promised
to pay the contracted subsidiaries. Once a new transmission is developed by the U.K.
subsidiary, the German FSA owner contracts the French subsidiary to manufacture the
transmissions and other foreign subsidiaries to distribute it. In effect, FSA ownership
shifts risk within the firm, in this case development risk, from the UK subsidiary to the
German subsidiary. It also shifts the rewards from innovative efforts. This creates the
classical principle agent problem within the context of the multinational firm.
Although the popular press has focused on internal contracting and licensing
associated with tax haven subsidiaries, tax havens are a subset of MNC contract and
licensing relationships. In reality, the majority of MNCs contract and license their FSAs
between the parent and/or subsidiaries that are not tax havens. The following table shows
the proportion of subsidiaries in the dataset that own the rights to FSAs.
Table 13: Breakout of FSA Ownership by Subsidiary Type
FSA
Owners
FSA
Users Total
Operational Non-Tax Haven Subsidiaries 4.9% 86.3% 91.1%
Operational Tax Haven Subsidiaries 0.5% 8.0% 8.5%
Pure Tax Haven Subsidiaries 0.3%
0.3%
Total 5.7% 94.3% 100%
95
As shown in Table 13, approximately 5.7% of subsidiaries own economic rights to
MNC FSAs, and the remaining 94.3% percent either contract or license FSAs from FSA
owners. Approximately 8.5% of subsidiaries are located in OECD-classified tax haven
countries and perform operational activities, 0.5% are FSA owners. Finally, 0.3% of
subsidiaries are pure tax haven FSA owners. This means that 1) relatively few
subsidiaries within the firm own the economic rights to the FSAs they create, 2) most
FSA owners are not located in tax haven countries, and 3) while all MNCs have internal
FSA owners, the majority of MNCs do not have tax haven FSA owners.
Economic ownership of FSAs is a means of allocating authority and control within
the organization. The entities within the MNC that bear the risks associated with creating
the FSAs also receive the rewards. In a sense, the FSA owners are the “entrepreneurs” of
the MNC. They bear risk associated with the MNCs activities while insulating other
subsidiaries within the firm from risks. As owners of the MNC’s most important value
generating resources, the FSA owners are responsible for generating future FSAs within
the firm. Research on subsidiary mandates examines the determinants and consequences
of allocating to a subsidiary the responsibility of a product line or division. Economic
ownership within the firm goes one step further: FSA owners are the entrepreneurs of the
firm that own the rights to the firm’s key value generating assets. FSA owners bear the
risk of the MNCs operations and coordinate activity by licensing the knowledge assets
and contracting activities by other subsidiaries and/or the parent within the firm.
96
Theoretical Background 5.3
5.3.1 MNCs and Innovation
MNC growth and profitability are based on the continuous process of generating,
developing, and implementing innovation (Buckley and Casson, 1976). Innovation is the
generation of a new idea, product, service, process, technology, or management practice
by an organization (Damanpour, 1991). While innovation occurs at the individual level,
subsidiary-level innovation is the aggregation of innovations of individuals employed at
the subsidiary location. This paper focuses on technological innovation. The literature
on the MNC broadly conceptualizes innovation to include technology, new products, new
management practices, and commercial applications of new knowledge (Buckley and
Casson, 2009). Research has shown innovation to be crucial for increasing firm market
share, profitability, and growth (Banbury and Mitchell, 1995; Cottrell and Nault, 2004;
Nerkar and Roberts, 2004; Morck and Yeung, 1991).
Early research on MNC innovation (e.g. Hymer 1976; Rugman 1981) viewed any
foreign R&D activities as ‘home-base exploiting’ (Kuemmerle, 1999) or ‘competence
exploiting’ (Cantwell, 1987), where foreign subsidiaries adapt FSAs developed by the
parent to the local host market. In the 1990’s, researchers began to view the MNC as a
network of subsidiaries that obtain and create knowledge in the local environment, and
share the knowledge across the network of subsidiaries (e.g. Bartlett and Ghoshal, 1989;
Nobel and Birkinshaw, 1998; Hedlund, 1994; Hakanson and Nobel, 1993). MNCs
establish foreign subsidiaries to draw on the local skills, knowledge, and expertise
necessary to generate FSAs. Subsidiary capabilities are generated based on the
competitive advantage of the region (Frost et al, 2002). Empirical research has shown
97
that foreign subsidiaries hold an important role in product development through building
technological capabilities and absorbing local technologies and knowledge (Pearce and
Papanastassiou, 1999; Zander, 1997; Driffield and Love, 2003).
5.3.2 Difficulties of Managing Innovation
Although the MNC has access to a diverse pool of knowledge and capabilities
from its foreign subsidiaries (Zhao, 2006; Birkinshaw, 1997; Kuemmerle, 1999; Lou,
2002), MNCs are faced with the challenge of fostering innovation across their
geographically dispersed network. Innovation is inherently difficult to monitor and
control since it involves combining implicit and explicit knowledge (Osterloh and Frey,
2000) and the outcome is uncertain. Innovation requires the investment of human capital.
For innovative activities, it can be very difficult to determine whether new knowledge has
been created or whether the effort has been invested for its creation. While new
initiatives that fail are observable, the lack of action on new ideas, concepts, or solutions
is unobservable. Consequently, it is difficult to distinguish between effort and luck. As a
result, monitoring mechanisms are ill suited for innovation activities.
Investing in R&D projects can be risky and costly. Subsidiaries may not want to
devote resources to developing the MNC’s FSAs unless they are adequately
compensated. Although the parent may finance R&D activities, in such cases the
subsidiary lacks control. The incentive to invest in innovation is muted when the
knowledge created is a public good (Holmstrom and Roberts, 1998). If FSAs are pure
public goods within the firm, subsidiaries may free ride on the efforts of others, and
conflicts can emerge over the use and control of the innovations created.
98
5.3.3 Property Rights Theory and Incentives
Grossman and Hart (1986) maintain that when activities are difficult to monitor
and control, asset ownership should be allocated to the entity whose contribution has the
largest marginal effect on the asset’s value. Asset ownership affects the incentives to
invest in human capital and knowledge creation (Wang et al, 2009). Because the owner
receives the excess income from the asset, ownership affects ex-ante investment
decisions (Grossman and Hart, 1986; Hart and Moore, 1990). When the asset owner is
not the entity that develops the resource, the developer has less incentive to devote the
full effort for maximizing value creation (Grossman and Hart, 1986).
Ownership provides the owners with power and control over the operations (Hart
and Moore 1990; Rajan and Zingales, 2000). The FSA owners coordinate and allocate
tasks to the non-owners. Rajan and Zingales (2000) maintain that the owners have the
power to punish the non-owners by withholding resource allocations, re-directing tasks,
and exiting businesses. The owners of the residual rights can also influence firm strategy.
Grossman and Hart (1986) provide the example of a coal boiler that has difficulty
processing impure coal. If the coal plant owner owns the coal mine, then it will direct
that the coal mine obtain better coal. In contrast, if the coal mine owns the coal plant, it
will direct that the plant technologically improve the boiler to handle the impure coal.
Ownership allows control over the solutions to problems and over the strategic direction
of future investments.
99
Hypotheses 5.4
FSA ownership provides a dual mechanism for incentivizing subsidiary
innovation: 1) the ability to appropriate income from the innovations created, and 2)
private benefits of control.
Ownership of MNC FSAs will determine the incentives of the subsidiaries to
invest in MNC FSAs. Ownership encourages more relationship-specific investments
(Grossman and Hart, 1986). Since the FSA owners bear the losses associated with failed
attempts to innovate, the subsidiaries that own FSAs have a vested interest to ensure
success. For subsidiaries, increased profitability from ownership of successful projects
makes it easier to justify bonuses and salary increases. Budd, Konings, and Slaughter
(2005) note that entity profitability can help to justify large wage payments to
management and workers. By assigning FSA ownership to entities within the MNC that
have the skills, knowledge, resources, expertise, and a better understanding of the risk,
the MNC can leverage human and physical resources to enhance its ability to innovate.
The second benefit of ownership is that it confers control rights (Hart and Moore,
1990). The FSA owner has control over the strategic decisions and how to allocate the
resources related to the asset. Research on subsidiary autonomy has long held that
control increases incentives and initiative (Birkinshaw, Hood, and Jonsson, 1998). At the
subsidiary level, Mudambi et al. (2007) examine the role of self-determination of
sourcing, hiring, marketing, and product development, on knowledge generation and find
that self-determination is positively associated with subsidiary innovation. Control of
knowledge creates incentives to invest in R&D (Levin et al., 1987). In a study of market
100
contracting relationships, Leiponen (2008) finds that the ability to control intangible
assets is positively associated with innovation.
In contrast, the control rights of the FSA owners blunt the entrepreneurial
incentives of the subsidiaries that are FSA users. When entities do not bear the full
effects of their decisions, they tend to underinvest due to the dissipation of benefits from
their investment (Fama and Jensen, 1983; Hart and Moore, 1990). Although the FSA
owner can delegate autonomy and decision making authority, the ability of the FSA
owner to intervene in management decisions reduces manager incentives (Aghion and
Tirole, 1997). Rotemberg and Saloner (1994) and Hart (2001) suggest that the ability of
residual rights owners to choose which ideas are implemented and to capitalize on the
benefits of the idea reduces managers’ incentives to produce valuable ideas. Stein (2001)
notes that it will discourage the entity from taking costly, non-contractible actions to
increase firm value. When the residual rights owner intervenes too often, it can stifle the
entity’s initiative (Aghion and Tirole, 1997). The owners of the control rights cannot
establish credible commitments to not intervene in delegated decision-making (Foss,
2003). Foss (2003) finds that the intervention by managers in decisions delegated to
employees dramatically reduces employee motivation. In a study on the allocation of
control rights in inter-firm exchanges of service firms, Leiponen (2008) finds that service
firms that yield control rights to clients are 20-30 percent less likely to introduce new
services.
Subsidiary FSA ownership, through the combination of the ability to appropriate
the income and control rights to the innovations they create, provides a strong incentive
for subsidiaries to innovate.
101
Hypothesis 1: Subsidiary FSA ownership will be positively associated with
innovation.
5.4.1 Contracting Relationships
The vast majority of subsidiaries are FSA users contracted by FSA owners to
perform activities within the firm. Moreover, many FSA owners contract and license
FSAs with other FSA owners. This section examines the question of whether the
contracting relationships matter for subsidiary innovation.
5.4.2 Tax Haven FSA Ownership, Contracting, and Subsidiary Innovation
The ownership of assets in an exchange relationship should be held by the entity
whose investment is the most important to the generation of future value. Asset owners
over invest in the asset because they receive all of the income from any assets created
from the investment (Grossman and Hart, 1986). The non-owning entities are likely to
underinvest since they do not receive any additional benefits, over the contractual return,
from any assets or income created in the exchange. Firms will choose an optimal
ownership structure that minimizes the overall loss in surplus due to investment
distortions (Grossman and Hart, 1986). Foss and Foss (2001) propose that costly
monitoring or verification will influence who will own an asset. Particularly when non-
contractible relationship-specific investments are required, ownership should be assigned
to the entities important to the development of the asset. Consequently, FSA ownership
may be misaligned when an entity within the firm owns the FSAs but does not directly
create the MNC FSAs.
102
The misalignment of ownership can reduce the incentives of the contracting
subsidiaries to invest in the asset. Roth and O’Donnell (1996) suggest that the
psychological identification of an agent with the principal is important in determining the
agency problem for global operations. Similarity of functions and reciprocity of
knowledge exchanges facilitate greater psychological identification between units. The
incentive effect on innovation activities will be particularly strong if the FSA owner does
not perform core activities. Gertner et al. (1994) note that assigning control rights to the
capital providers in an internal capital market may be costly since it diminishes
managerial incentives. Pure tax haven entities do not perform operational activities nor
do they trade with other entities. If the FSA owner performs activities that contribute to
the generation of the FSA, the FSA users may more readily concede that the cash flow to
the FSA owner is warranted. In contrast, if the FSA owner does not perform core
activities, FSA users may become disenchanted since the fruits of their labor are captured
by an entity whose activities do not directly contribute to the development of the FSAs.
The literature on internal capital markets discusses the “dark side” of resource
reallocation within the firm. Resource reallocation reduces incentives. When units differ
in resources and investment potential, resources are reallocated despite manager efforts
(Inderst and Laux, 2005). Brusco and Panunzi (2005) argue that removing the residual
income generated by a unit reduces manager incentives to generate cash flows. This
leads to smaller overall value creation and thus creates less residual income to reinvest
(Brusco and Panunzi, 2005).
Hypothesis 2: Contracting with a Pure Tax Haven FSA owner will be negatively
associated with subsidiary innovation.
103
5.4.3 FSA Ownership, Contracting with Parent, and Innovation
Research on headquarter-subsidiary relations has long held that connections with
headquarters are important. The subsidiaries that contract with parent FSA owners have
direct, ongoing operational ties with the parent. These ties involve routine reporting,
regular invoices sent to the parent for activities performed, communication about
budgets, projects, and changes in operations. Managers of the contracting subsidiaries
have higher visibility with corporate headquarters, which can provide incentives of career
advancement and greater access to resources if the subsidiary performs well. Interunit
communication is important for the creation of innovations (Burns and Stalker, 1961).
Prior research has shown that communication and positive attention from the parent
increases subsidiary initiative (Birkinshaw, 1999; Bouquet and Birkinshaw, 2008).
Additionally, interactions with the parent can foster shared identity, which has
been considered necessary for motivating subsidiaries to innovate (Ghoshal and Bartlett,
1988). By sharing strategy, goals, and values with the corporate group, subsidiaries are
more likely to better understand their role in the organization and may be more
accommodating to the needs of other units, as well as motivated towards innovating for
the group (Ghoshal and Bartlett, 1988). Moreover, shared organizational identity
incentivizes knowledge sharing, which enhances innovation. Ghoshal and Bartlett (1988)
find that high levels of headquarter-subsidiary communications are positively associated
with subsidiary creation, adoption and diffusion of innovations.
Incentives for recognition, shared identity with the corporate group, and
knowledge sharing can work together to increase innovation.
104
Hypothesis 3: Contracting with the Parent will be positively associated with
innovation.
Methods 5.5
The hypotheses were tested using a unique, confidential, subsidiary-level dataset
constructed from several sources. Information on MNC internal FSA ownership was
compiled from MNC transfer pricing reports from a consulting firm. In compliance with
OECD guidelines and government regulations, MNCs are required to document their
intra-firm transactions each year. The transfer pricing reports contain detailed
information on the contractual relationships between subsidiaries, including the functions
performed and risks borne by the affiliates, as well as the economic owner(s) of the FSAs
in the exchange relationship. From the transfer pricing data, I coded each subsidiary’s
functions, whether or not the subsidiary was an FSA owner, and the FSA owners with
which the subsidiary contracts. Subsidiary financial data was pulled from Bureau Van
Dijk’s Orbis database. Patent data was collected from the USPTO since it provides the
location of each inventor on the patent, making it possible to match patents to inventor-
subsidiary locations. Data for the revealed technological advantage index came from the
OECD Patent database, based on the inventor’s country location and patent industry.
Finally, market concentration data came from Compustat’s Global Vantage database.
For subsidiary patents, I searched for all granted patents assigned to each MNC or
to any subsidiary within the MNC’s group. Each patent was matched to subsidiary(ies)
based on inventor city, state, and country information and to a year based on the filing
date of the patent. A total of 50,934 patents were patented by the MNCs in the sample
over the sample period, of which 29,028 had only one inventor and 10,711 patents had
105
multiple inventors from the same subsidiary location and therefore were coded to only
one subsidiary. The remaining 11,195 patents had inventors from more than one
subsidiary location and were therefore coded to multiple subsidiaries. On average,
subsidiaries in the sample applied for 0.46 patents each year.
Relying on the USPTO for patent data can lead to a bias towards U.S. MNCs and
U.S. subsidiaries. However, firms that are active in the U.S. have a strong incentive to
file for intellectual property protection in the United States. As 72 percent of the MNCs
in the dataset are U.S.-headquartered and all MNCs in the dataset have at least one
location in the U.S., the bias should be somewhat mitigated by the fact that all MNCs in
the sample have incentive to patent their innovations with the USPTO.
Although the sample size varies depending on the analysis, the following provides
an overview of the construction of the base sample. The starting sample was composed
of 28,837 subsidiary-year observations on 7,156 subsidiaries from 95 MNCs, over the
1997 through 2011 time period. The MNCs in the sample are large firms – the average
number of subsidiaries per MNC is 183. However, the mean is skewed. The number of
subsidiaries per MNC ranges from 1 to over 800 subsidiaries. All subsidiary types (e.g.
R&D, distribution, manufacturing, financial, etc.) are included in the sample since FSA
ownership may be allocated to any subsidiary and any subsidiary may patent an
invention. Merging in the market concentration index reduced the sample by 4,496
observations to 24,341 subsidiary-year observations. Missing MNC diversification data
further reduced the sample by 1,184 observations to 23,157 subsidiary-year observations
on 5,919 subsidiaries. In order to make the regressions with two-digit SIC industry fixed
effects estimable, subsidiaries in two digit industries with zero patenting were excluded
106
from the sample, which reduced the sample to 21,148 observations. The final starting
sample was composed of 5,404 subsidiaries and 21,148 subsidiary-year observations.
For the regressions incorporating subsidiary-specific financial data, Orbis had
financial information on 9,987 subsidiary-year observations. Missing data were entered
from the company consolidating financial statements used for the transfer pricing reports
to avoid the unnecessary loss of observations. Consolidating financial data were used for
9,579 observations, leading to a total of 19,566 subsidiary-year observations. Including
lagged subsidiary-level financial variables reduces the sample by 6,657 observations on
4,274 subsidiaries and 14,491 subsidiary-year observations.
5.5.1 Variables
Dependent Variable
Subsidiary Innovation. Consistent with prior research (e.g. Phene and Alemeida,
2008; Ahuja, 2000; Rothaermel and Hess, 2007), total patents was used as a measure of
innovation. Total Patents was measured as the total number of successful patent
applications associated with the subsidiary’s inventors in each year. The patent filing
date was used to match the patents to years since it is closer to the timing of the invention
than the patent grant date.
Independent Variables
FSA Owner is a binary indicator equal to one if the subsidiary owns the economic
rights to the MNC’s FSAs, and zero otherwise. This measure was coded based on the
information contained in the transfer pricing reports. The transfer pricing reports
107
explicitly identify the subsidiaries that are the FSA owners (See Chapter 3 for a detailed
description of the coding).1
Contract with Pure Tax Haven FSA Owner is a binary indicator set equal to 1 if
the subsidiary is contracted by or licenses FSAs from a pure tax haven FSA owner. An
FSA owner was coded as a pure tax haven FSA owner if it was located in tax haven
country and does not perform R&D, manufacturing, distribution, sales and marketing
activities, or financial trading (for banking firms). Approximately 21 percent of the
observations in the sample contract with pure tax haven FSA owners.
Contracts with Parent FSA Owner is a binary indicator set equal to one if the
subsidiary is contracted by the parent or licenses FSAs from the parent. Approximately
25 percent of the observations in the sample contract with a parent FSA owner.
Control Variables
Revealed Technological Advantage. Host country specialization in industry
domains affects the generation of new ideas and access to skilled and highly capable
workforce (Kogut and Chang, 1991; Anand and Kogut, 1997). Subsidiaries assimilate
knowledge from their host-country local environment (Kim, 1997; Kummerle, 1996;
Westney, 1992). I controlled for country-industry specialization using the Revealed
Technological Advantage index, calculated as the country’s share of patents in the focal
subsidiary’s industry divided by the country’s share of patents in all industries (OECD,
2011). The index equals zero when the country has no patents in the given industry, one
if the country has no specialization, and is greater than one if the country is specialized in
1 Ownership is also supported by real flows within the MNC, and legal, written contracts between MNC
entities.
108
the industry. The mean value of this variable was 0.90. Revealed technological
advantage is expected to be positively associated with innovation.
Market Concentration. Local market competition drives the need to innovate and
is associated with high patenting activities. I control for each subsidiary’s country-
industry market competition using the market concentration index, calculated as the sum
of the portion of total industry revenues earned by the four largest firms in the industry
and country location of the subsidiary. This variable equals zero for highly competitive
markets and increases for less competitive markets. Low values represent greater market
competition, thus it is expected that this variable will be negatively associated with
innovation.
Change in Country Effective Tax Rate. In an exploratory analysis of changes in
ownership, the change in the host country effective tax rate is used as a control. The host
country effective tax rate is calculated as the total taxes paid by firms in the country,
divided by the total profit before tax of firms in the country. The change in country
effective tax rate was calculated as the current period effective tax rate minus the prior
period effective tax rate, and divided by the prior period effective tax rate. It is expected
that MNCs will transfer FSA ownership away from a subsidiary if there is a large
increase in the country effective tax rate.
GDP Growth. Since changes in FSA ownership may be due to economic
conditions, GDP Growth is included as a control using the change in gross domestic
product for each country. GDP growth is expected to be negatively associated with
changes in ownership.
109
MNC Size. Innovation can be affected by the resources available for innovation
activities. Large MNCs can draw on a large set of resources and exploit scale economies
in R&D (Cohen, 2010). I control for MNC Size using as the natural log of total MNC
revenues. I rely on revenues as it is less affected by MNC strategic decisions as
investment in assets and R&D expenditures.2 I expect MNC size to be positively
associated with innovation.
MNC Average Subsidiary Size.3 It is expected that MNCs that have larger
subsidiaries will be more likely to grant FSA ownership to their subsidiaries. MNC
Subsidiary size is calculated as total MNC revenues divided by the total number of MNC
subsidiaries.
MNC Diversification. By engaging in diverse business segments, MNCs can take
advantage of economies of scope for their R&D activities (Cohen, 2010). A total entropy
diversification measure is used to capture the diversity of a firm’s activities (e.g. Bowen
and Wiersema, 2005; Palepu 1985). The variable is calculated as follows:
Total Entropy = ∑
Si represents the MNC’s share of total sales in business segment i. N represents
the total number of business segments in which the MNC operates. The measure is
calculated using the business segment information from Orbis and, in the case where
Orbis data was unavailable, from the consolidated financial data in the transfer pricing
2 In results not reported herein, I also estimated alternative specifications with MNC R&D intensity,
calculated as MNC R&D expenditures divided by MNC total revenues, R&D stock, calculated using the
perpetual inventory method of Rt+(1-d)Rt-1 where d=.15 (Hall et al 2005), and MNC assets. The
alternative analyses yielded results consistent with those contained in Table 19.
3 Since subsidiary size is directly affected by FSA ownership and is highly correlated with other variables
in the analysis, subsidiary size is excluded as a control.
110
reports. This variable equals zero for single business firms and increases with greater
levels of diversification.
Subsidiary R&D Intensity. R&D expenditures represent inputs into the innovation
process. Prior research suggests that R&D intensity is significantly associated with
patents (Hall, Griliches, & Hausman, 1986; Mueller, 1966). Subsidiary R&D Intensity is
measured as the focal subsidiary’s total R&D expenditures divided by its total revenue.
Subsidiaries in the sample had an average R&D Intensity of 4.0 percent.
Subsidiary Role. Subsidiaries that perform more activities tend to be more
innovative. Consistent with prior research (e.g. Fey and Furu, 2008; Almeida and Phene,
2004), I control for subsidiary role using an index of the number of activities that the
subsidiary performs. The index is a count variable, giving a weighting of one to each of
the four functions: R&D, manufacturing, distribution, and service. The index has a value
of one if the subsidiary only performs one function and a value of four if the subsidiary
performs all four functions.
M&A. M&A is a dummy variable, equal to one if the subsidiary was acquired as
part of a merger or acquisition, and 0 if otherwise. I expect this variable to be positively
associated with FSA ownership as well as positively associated with innovation.
Industry and Year Effects. I control for industry using two-digit Standard
Industrial Classification (SIC) code fixed effects and differences in time using year fixed
effects. All independent variables are lagged by one year.
5.5.2 Estimation
A primary concern in conducting the analysis is whether the relationship between
FSA ownership and innovation is reversed. MNCs theoretically should assign ownership
111
to their most capable, value generating subsidiaries. This would produce an upward bias
in the coefficient for FSA ownership in the results. Disentangling the endogenous
decision for a particular subsidiary to own an FSA from the effects of ownership on
innovation can be very difficult. To tackle the endogeneity concerns, multiple
approaches are used.
First, a negative binomial model is estimated with a pre-sample mean scaling fixed
effect to control for subsidiary-level unobserved heterogeneity in innovation (Blundell,
Griffith, and Van Reenen, 1999). Since there is a long history of pre-sample patenting
available from the USPTO (from 1975 onwards), the pre-sample average number of
patents produced by the subsidiary can be used as an initial condition to proxy for
unobserved differences, such as subsidiary capabilities, that can affect innovation
(Blundell et al. 1999; Aghion, Van Reenen, and Zingales, 2013). Second, I use
propensity score matching and match each FSA owner to a non-owning subsidiary in the
same country, two-digit SIC industry, and year, and with similar subsidiary-level and
MNC-level characteristics. For robustness, I compare the results to an instrumental
variables analysis. Finally, I explore the effects of removing ownership from a subsidiary
on that subsidiary’s innovation output. Each approach has its advantages and drawbacks.
For the baseline specification, I rely on a negative binomial model since the
dependent variable is a count variable characterized by overdispersion. The baseline
specification takes the following form:
E(Pit | Xit) = exp{β1Xi(t-1)+ β2Zi(t-1)+φj +τt }, (1)
112
Where X is the set of hypothesized independent variables lagged by one year, Z is
a vector of lagged control variables, φj is a set of 2-digit industry controls, and τt is a set
of year fixed effects.
For the initial conditions specification, I include the pre-sample average number of
patents for each subsidiary (Blundell et al, 1999), calculated as the sum of successful
patents applied for over the pre-sample period, divided by the total number of years from
the first patent filing to 1996, the year before entering the sample. An advantage to using
initial conditions is that it is able to control for time-invariant, unobserved differences in
the innovative capabilities of subsidiaries. However, FSA ownership structures tend to
be stable over time, with many subsidiaries holding ownership for decades (less than one
percent of observations change in the sample). As a result, the pre-sample average
measure may pick up part of the effect of FSA ownership and create a downward bias in
the results.
The second analysis applies propensity score matching (Rosenbaum and Rubin,
1983), which has been used extensively in economics and management research.
Propensity score matching attempts to uncover the effect of the “treatment” – in this case,
FSA ownership – by comparing the group of treated firms to a control group of matched
similar firms. I match each FSA owner to a non-FSA owning subsidiary with similar
characteristics in the same country, two-digit industry, and year. Subsidiaries were
matched based on whether they were located in the same continent as their parent, the
average size of the MNC’s subsidiaries, whether or not the subsidiary was acquired,
subsidiary R&D intensity, and subsidiary role. Matching within country, industry, and
year removes the effects of unobserved country, industry and year-specific factors. The
113
first stage of the propensity score analysis relies on a logit to predict FSA ownership
based on the subsidiary characteristics, and country, industry, and year matching criteria.
An algorithm is then used to match FSA owners to FSA users based on their predicted
probabilities from the logit regression. I used nearest neighbor matching, without
replacement. Only those observations on common support were included in the analysis.
In order to produce unbiased results, the characteristics of the control subsidiaries
should not be significantly different from the characteristics of the treatment subsidiaries.
Table 14 contains the results of balancing tests, which indicate that the variables
were balanced after matching. A t-test of the difference between the treated and control
groups’ patenting behavior was significant at p<.001.
Table 14: Balancing Test for Propensity Score Matched Sample
Variable Treated Control t p>t
Same Continent Unmatched 0.12 0.29 -14.19 0.000
Matched 0.12 0.13 96.10 -0.55 0.583
MNC Avg Subsidiary Size Unmatched 0.48 0.32 8.08 0.000
Matched 0.48 0.48 97.10 0.13 0.900
M&A Unmatched 0.20 0.23 -3.18 0.001
Matched 0.20 0.20 90.90 -0.23 0.821
R&D Intensity Unmatched 0.05 0.03 3.53 0.000
Matched 0.05 0.05 90.20 0.17 0.862
Role Unmatched 1.41 1.17 13.61 0.000
Matched 1.41 1.41 99.70 0.02 0.981
Revealed Technological Advantage Unmatched 0.86 0.82 2.85 0.004
Matched 0.86 0.86 97.50 0.05 0.960
Country Effective Tax Unmatched 0.22 0.23 -6.71 0.000
Matched 0.22 0.22 85.90 0.73 0.468
Market Concentration Unmatched 0.98 0.97 4.82 0.000
Matched 0.98 0.98 74.80 -1.49 0.137
Mean % Bias
Reduction
t-test
114
An advantage to the propensity score method is that it is able to control for
unobserved heterogeneity based on observed characteristics. A limitation to this
approach is that to the extent that there are unobserved factors uncorrelated with the
controlled factors, the results can be biased. As robustness check, I also estimated
instrumental variables analysis using generalized method of moments for negative
binomial models and also a two stage least squares.
In an attempt to uncover the effects of a change in ownership, I exploit the 24
changes in FSA ownership in the sample. Two separate analyses are conducted. First, on
the sample of subsidiaries that sold their FSAs, I estimate the change in innovation post-
FSA ownership change. Second, I use a propensity score estimator to reweight
subsidiaries in the matched sample for the probability of having ownership removed.
Each FSA owner whose FSAs are sold to another MNC entity is matched to a “control”
FSA owner with similar characteristics in the year of change. Once the matched FSA
owner is identified, I pool all years of data on the matched control and treated
observations and estimate the probability that the subsidiary has FSA ownership
transferred away. The estimated probability is the propensity score. The characteristics
used to obtain the propensity score are the change in country effective tax rate, lagged
R&D intensity, lagged GDP growth, and country, 2-digit industry and year exact
matching. The analysis is restricted to firms that fall within common support. The
following table contains the t-tests for the differences in means of the unmatched and
matched samples.
115
Table 15: Balancing Test for Change in FSA Ownership Matched Sample
Once the matched sample is obtained, the propensity score estimates are
transformed into weights (Dehejia and Wahba, 1999; Busso, DiNardo, and McCrary,
2009), by weighting each treated firm by 1/ ̂ and weighting each control by 1/(1- ̂). This
creates an estimate of the Average Treatment Effect of transferring ownership away from
the subsidiary.
Busso, DiNardo, and McCrary (2009) note that the reweighting technique has
better small sample properties than using simple propensity score matching. The analysis
Variable Treated Control t p>t
Change in Country Effective Tax Rate Unmatched 0.01 -0.01 1.48 0.141
Matched 0.01 0.03 35.70 -0.55 0.584
GDP Growth Unmatched 3.19 1.69 2.31 0.022
Matched 3.19 2.95 83.70 0.26 0.796
Market Concentration Index Unmatched 0.92 0.94 -0.45 0.651
Matched 0.92 0.92 87.10 -0.03 0.974
Subsidiary Size Unmatched 7.61 6.86 1.18 0.240
Matched 7.61 7.16 40.30 0.46 0.646
MNC Size Unmatched 13.12 13.72 -1.20 0.230
Matched 13.12 13.18 88.90 -0.11 0.913
MNC Subsidiary Size Unmatched 0.78 1.08 -0.66 0.510
Matched 0.78 1.18 -33.00 -0.90 0.374
R&D Intensity Unmatched 0.16 0.18 -0.27 0.787
Matched 0.16 0.09 -128.70 1.01 0.321
Inititial Patents Unmatched 0.69 1.07 -1.31 0.190
Matched 0.69 1.03 8.70 -1.22 0.231
Unmatched -0.25 -0.11 -0.72 0.473
Matched -0.25 -0.12 5.40 -0.79 0.438
Unmatched -0.12 -0.03 -0.45 0.652
Matched -0.12 0.01 -45.40 -0.66 0.512
Mean t-test% Bias
Reduction
Change in Patents Year Prior to Year
of Treatment
Lagged Change in Patents - Two
Years Prior to Year Before Treatment
116
allows one to control not just based on time-invariant characteristics, but also for time
varying characteristics that affect selection. I estimate a negative binomial model to test
the effects of transferring ownership away from the subsidiary unit on the reweighted
matched sample and cluster the standard errors by subsidiary. Since there are few
observations of change, this analysis is considered exploratory.
Results 5.6
5.6.1 Descriptive Statistics
The following tables contain descriptive statistics and correlations for the variables
used in the analyses.
Table 16: Descriptive Statistics and Correlations for Initial Conditions Analysis
Variables Mean s.d. 1 2 3 4 5 6 7 8 9 10 11
1 Total Patents 0.38 5.20
2 FSA Owner Dummy 0.07 0.25 0.12
3 Revealed Technological Advantage 0.90 0.50 0.05 0.07
4 Market Concentration Index 0.58 0.07 -0.04 0.03 -0.05
5 MNC Size 15.69 1.81 0.07 0.03 -0.04 0.20
6 MNC Diversification 0.52 0.56 0.06 -0.07 0.04 0.09 0.39
7 Initial Conditions 0.09 0.39 0.62 0.17 0.09 -0.03 0.09 0.07
8 M&A 0.29 0.46 -0.01 0.01 0.06 0.02 0.05 0.17 0.01
9 R&D Intensity 0.04 0.20 0.06 -0.02 0.04 -0.09 -0.11 -0.04 0.13 0.03
10 Role 1.17 0.62 0.16 0.14 0.03 0.00 0.02 0.00 0.26 -0.04 0.20
11 Contracts with Parent FSA Owner 0.25 0.43 0.01 -0.11 -0.03 -0.17 -0.57 -0.21 0.05 -0.17 0.12 0.08
12 Contracts with Pure Tax FSA Owner 0.21 0.41 -0.05 -0.11 -0.03 -0.01 -0.16 0.01 -0.08 -0.08 0.02 -0.06 0.05
Number of observations=1621. Year=2006. Correlations greater than or less than 0.02 are significant at *p<.05.
117
Table 17: Descriptive Statistics and Correlations for Propensity Score Analysis
On average, subsidiaries in the sample patent less than one patent a year. However, this
variable is skewed, with a long right hand tail. Approximately seven percent of
observations in the sample are FSA owners. One quarter of subsidiaries contract with
parent FSA owners and 21 percent contract with pure tax haven FSA owners. MNC
entities can contract with multiple FSA owners. Although not shown here, 55 percent of
subsidiaries contract with operational subsidiary FSA owners and 15 percent contract
with operational tax haven FSA owners. Turning to the correlations, as expected, the
initial conditions control is highly correlated with patenting (r=0.62). The largest
correlation between independent variables is between contracting with parent and MNC
size (r=-.57). Smaller firms have a larger proportion of subsidiaries contracting with the
parent. Because of the high correlation, the baseline specification for testing the
hypotheses relies on MNC diversification instead of MNC size.
5.6.2 Changes in FSA Ownership
I explore the effects of change in FSA ownership on innovation. There were 24
observations of change in the sample. As summarized in the table below, one quarter of
Variables Mean s.d. 1 2 3 4 5 6 7 8 9 10
1 Total Patents 1.17 6.52
2 FSA Owner Dummy 0.49 0.50 0.14
3 Revealed Technological Advantage 0.90 0.60 0.08 0.05
4 Market Concentration Index 0.59 0.03 -0.13 -0.01 -0.22
5 MNC Size 15.73 1.79 0.08 -0.07 -0.15 0.25
6 MNC Diversification 0.46 0.50 0.04 -0.11 0.02 0.05 0.27
7 M&A 0.22 0.42 -0.04 0.00 -0.01 0.00 0.01 0.25
8 R&D Intensity 0.06 0.31 0.03 -0.10 0.07 -0.16 -0.05 0.02 0.08
9 Role 1.48 0.78 0.22 0.02 0.05 -0.13 -0.15 0.05 -0.02 0.15
10 Contracts with Parent FSA Owner 0.14 0.35 0.01 -0.12 -0.05 -0.03 -0.40 -0.08 -0.13 0.01 0.14
11 Contracts with Pure Tax FSA Owner 0.09 0.29 -0.06 -0.13 -0.09 0.08 -0.21 0.01 -0.01 -0.01 -0.13 0.19
Number of observations=315. Year=2006.
118
the changes transferred FSA ownership away from the parent, a little less than half
transferred FSA ownership away from operational subsidiaries, and approximately one
third transferred it away from tax haven units.
Table 18: Subsidiary FSA Ownership Change
Approximately 24% of the changes were transferring ownership to operational
non-tax haven subsidiaries, another 24% transferred ownership to operational tax haven
subsidiaries, 36% were transferred to pure tax haven subsidiaries, and 16% were
transferred to the parent.
The figures below display the mean number of patents and patent citations
received on those patents, respectively, for the subsidiaries over the three years pre- and
post-change.
Observations % of Total
Total Number of Changes 24
FSA Ownership Transferred Away From:
Parent 6 25%
Operational Subsidiary 11 46%
Operational Tax Haven 5 21%
Pure Tax Haven 2 8%
119
Figure 11: Average Patenting and Citations of Patents for FSA Ownership Change
The data shows a marked change in the trend of patenting and patent quality in the
years after ownership is transferred away from the subsidiary. The mean number of
patents increases over time, but drops in the year that ownership is removed from the
subsidiary and continues to drop after the change. The average number of citations on
patents issued three years prior to and three years post change indicate a slight drop in
quality of patents in the year of change before realizing a steep drop in the years after the
change.
0
1
2
3
4
5
6
7
t-3 t-2 t-1 t=0 t+1 t+2 t+3
Total Patents
Total Patents
0
5
10
15
20
25
30
t-3 t-2 t-1 t=0 t+1 t+2 t+3
Average Citations
Average Citations
120
5.6.3 Regression Results
Table 19: Random Effects Negative Binomial Results
Table 19 contains the results for the initial conditions and propensity score
matching analyses. Column 1 contains the basic initial conditions model, Columns 2-3
add the independent variables Contract with Pure Tax Haven and Contract with Parent,
respectively. Column 4 contains all three hypothesized independent variables. Column 5
adds the lagged endogenous subsidiary-level variables R&D Intensity and Role, which
(1) (2) (3) (4) (5) (6) (7) (8) (9)
FSA Owner Dummy 0.66*** 0.64*** 0.77*** 0.75*** 0.42** 0.39* 0.85*** 0.83** 0.85**
(0.15) (0.15) (0.15) (0.15) (0.16) (0.15) (0.25) (0.28) (0.26)
Contract with Pure Tax Haven -0.32* -0.31* -0.37* -0.39** -1.81** -1.03*
(0.15) (0.15) (0.16) (0.15) (0.56) (0.52)
Contract with Parent 0.55*** 0.55*** 0.22+ 0.04 0.52 0.28
(0.11) (0.11) (0.12) (0.12) (0.34) (0.33)
Revealed Tech. Advantage -0.02 -0.02 0.01 0.01 0.08 0.09 1.06*** 1.26*** 1.09***
(0.09) (0.09) (0.09) (0.09) (0.09) (0.09) (0.26) (0.28) (0.26)
Market Concentration -1.05*** -1.05*** -0.99*** -0.99*** -0.60* -0.60* -0.74 -1.29 -0.77
(0.27) (0.27) (0.27) (0.27) (0.31) (0.30) (0.76) (0.83) (0.77)
Initial Conditions 2.87*** 2.86*** 2.80*** 2.78*** 2.37*** 2.47***
(0.10) (0.10) (0.10) (0.10) (0.13) (0.14)
MNC Diversification 0.04 0.05 0.13 0.15 0.11 0.60* 0.56+ 0.67*
(0.09) (0.09) (0.10) (0.10) (0.11) (0.28) (0.30) (0.29)
MNC Size 0.03
(0.04)
M&A -0.16 -0.18 -0.08 -0.09 -0.13 -0.15 -0.11 0.02 -0.09
(0.11) (0.11) (0.11) (0.11) (0.13) (0.13) (0.29) (0.30) (0.29)
R&D Intensity 0.41*** 0.37*** 0.31 0.28
(0.10) (0.10) (0.22) (0.22)
Role 0.53*** 0.49*** 0.98*** 0.90***
(0.08) (0.08) (0.15) (0.16)
Industry Fixed Effects Yes Yes Yes Yes Yes Yes Yes Yes Yes
Year Fixed Effects Yes Yes Yes Yes Yes Yes Yes Yes Yes
Country Fixed Effects No No No No No No Yes Yes Yes
Constant -1.97*** -1.94*** -2.29*** -2.26*** -3.54*** -3.82*** -0.94 2.19 -0.95
(0.37) (0.37) (0.38) (0.38) (0.45) (0.74) (1.86) (1.89) (1.87)
Number of Observations 21148 21148 21148 21148 13744 14491 2490 2490 2490
Wald Chi Squared 1894*** 1912*** 1877*** 1895*** 1515*** 1544*** 208*** 174*** 211***
AIC 9321 9319 9299 9297 7528 7772 2428 2456 2427
Log Likelihood -4618 -4615 -4606 -4603 -3717 -3839 -1157 -1171 -1155
Initial Conditions
Propensity Score Matched
Sample
Robust, clustered standard errors are in parentheses. Two-tailed tests for variable coefficients. †p<.10; *p<.05; **p<.01; ***p<.001.
121
can be affected by FSA ownership. Column 6 replaces MNC diversification with MNC
size. Columns 7-9 contain the propensity score matched sample results. Because
propensity score matching is based on a matched sample, the number of observations
decreases from 21,148 to 2,490. Turning briefly to the model statistics, estimations
including all independent variables have better model fits, as indicated by the lower AIC
values.
For the control variables, as expected, the initial conditions control has a strong
positive association with patenting (p<.001, Columns 1-6). Including the initial
conditions control in the model causes Revealed Technological Advantage and MNC
Diversification to become insignificant. Both Revealed Technological Advantage and
MNC Diversification are positive and significant in the matched sample analysis (p<.001
and p<.05, Column 7). Since Revealed Technological Advantage and MNC
Diversification tend to be relatively stable over time, the initial conditions control may be
capturing part of the effect of these variables. Market concentration is negative and
significantly associated with innovation, although this effect is reduced once R&D
intensity and subsidiary role are added to the models. The negative and significant
relationship indicates that when there is less local industry market competition,
subsidiaries are less likely to innovate. Surprisingly, M&A is insignificant in both the
initial conditions model and the propensity score matching model. After controlling for
factors associated with innovation, it appears that acquired subsidiaries are no more likely
to innovate than greenfield subsidiaries. As expected, subsidiary R&D intensity and role
are positive and significantly associated with patenting (p<.001, Columns 5 and 6).
122
FSA ownership is positive and significantly associated with innovation across all
estimations. All columns report coefficients on FSA ownership between .39 and .85.
The results provide consistent support for Hypothesis 1. Hypothesis 2 is also supported:
Contract with Pure Tax Haven FSA Owner is negative and significantly associated with
innovation. For Hypothesis 3, the initial conditions and propensity score models provide
mixed support. The coefficient for Contract with Parent FSA Owner is positive and
significant in Columns 3-4 (p<.001). However, it becomes insignificant once Subsidiary
R&D Intensity and Subsidiary Role are added to the model. Moreover, Contract with
Parent FSA Owner is insignificant for the estimations using the propensity score matched
sample (Columns 8-9, p>.05).
To further explore the effects of FSA ownership on innovation, I examine changes
in FSA ownership. The results to the exploratory analysis of change in FSA ownership
are shown in Table 20. There were 24 instances of change over the sample window.
Lagged variables and missing data reduced the number of change observations to 21 in
the initial conditions sample. The total subsidiary-year observations for subsidiaries that
had ownership removed was 106 in the final sample. Lagged values in the first stage of
the propensity score analysis reduced the number of change observations to 16 in the re-
weighted propensity score analysis. The pooled control and treated sample for the
reweighted propensity score analysis was 159 subsidiary-year observations.
123
Table 20: Results for Transferring Ownership of FSAs Away from Subsidiary
Columns 1-3 present the results for within-subsidiary change in patenting. A
negative binomial model is estimated on the sample of subsidiaries whose rights to FSAs
were sold to another unit within the MNC. The results to the reweighted propensity score
analysis are shown in Columns 4-6. Both analyses provide clear evidence of a drop in
innovation output. Across all six columns, the coefficient for post-change in ownership
(1) (2) (3) (4) (5) (6)
Post Change -0.96** -0.82* -1.44*** -1.32** -1.35** -1.32**
(0.33) (0.32) (0.36) (0.45) (0.46) (0.45)
Revealed Tech. Advantage 2.74 3.22+ 2.69 0.34
(1.99) (1.81) (1.80) (0.98)
Market Concentration -0.79 -0.63 -0.10 -8.74 -9.09 -8.74
(0.88) (0.90) (0.87) (7.70) (8.25) (7.71)
Initial Conditions 1.28+ 1.02 1.40*
(0.68) (0.68) (0.62)
MNC Diversification 0.89 1.15* 1.42*** 1.15*
(0.70) (0.51) (0.36) (0.51)
MNC Size 0.33 0.30
(0.22) (0.22)
R&D Intensity 0.48 0.11 1.53 1.75
(1.14) (0.98) (1.30) (1.78)
Role 0.86+ 0.40 0.40 0.41+ 0.43*
(0.45) (0.34) (0.41) (0.21) (0.21)
Industry Fixed Effects Yes Yes Yes Yes Yes Yes
Year Fixed Effects Yes Yes Yes Yes Yes Yes
Country Fixed Effects No No No Yes Yes Yes
Constant 2.89 3.74 8.89 7.44 7.73 7.44
(435.11) (553.44) (667.53) (6.84) (7.28) (6.83)
Number of Observations 106 102 100 159 157 159
Wald Chi Squared 171*** 175*** 141*** 360*** 469*** 346***
AIC 288 285 262 344 346 342
Log Likelihood -120 -118 -109 -150 -152 -150
Robust, clustered standard errors are in parentheses. Two-tailed tests for variable coefficients.
†p<.10; *p<.05; **p<.01; ***p<.001.
Subsidiaries with Change
Propensity Score Reweighted
Matched Sample
124
is negative and significant (p<.05), suggesting that subsidiaries that have FSA ownership
transferred away produce fewer innovations.
5.6.4 Robustness Tests
Several robustness tests were undertaken to check the sensitivity of the results.
First, instrumental variables analysis was estimated on the full sample using generalized
method of moments and two-stage least squares. Two variables were used as
instruments: 1) the percentage of FSA owners in each subsidiary’s one-digit SIC industry
and country location (correlation with FSA Owner Dummy 0.44), and 2) a binary
indicator for whether the subsidiary is located on the same continent as the parent
(correlation with FSA Owner Dummy 0.12). The instrumental variables analyses yielded
positive and significant results for FSA ownership (p<.001) and contracting with a parent
FSA owner (p<.001), and negative and significant results for contracting with a pure tax
haven FSA owner (p<.001). Second, alternative controls were included in the models
such as the natural log of subsidiary revenue and MNC number of subsidiaries instead of
MNC size. Third, variables were entered into the models one at a time. These alternative
analyses yielded consistent results for the hypothesized variables as those shown in the
tables above.
Discussion 5.7
The theory of the multinational firm emphasizes the role of firm-specific
advantages in the existence of the firm and as a source of above normal returns (e.g.
Dunning, 1977; Rugman and Verbeke, 2001). The free-flow of knowledge within the
firm and the ability to leverage FSAs across the MNC’s network of affiliates are held as
125
key advantages of the MNC. However, markets for knowledge exist inside MNCs. The
parent and subsidiaries buy and sell rights to the MNC FSAs. The transactions that occur
within the firm are not solely financial, they involve real activities and transfer risk,
rewards, and control within the MNC. A key finding of this research is that markets for
knowledge within the firm can enhance firm innovation.
This study draws on property rights theory to predict the effects of FSA ownership
on subsidiary innovation. Firms select organizational designs to cope with information
and knowledge problems (Holstrom and Roberts, 1998). Allocating ownership rights to
FSAs within the firm distinguishes which subsidiaries have control over the FSAs and
bear the risk and rewards of developing them. Ownership has a strong effect on the
incentives for investing by influencing the distribution of profits and control over the
asset and its future strategic uses (Grossman and Hart, 1986; Hart and Moore, 1990; Kim
and Mahoney, 2005; Whinston, 2001). I predict that while FSA ownership provides
incentives to invest in the creation of new FSAs, lack of ownership reduces such
incentives.
The dynamics of innovation within the MNC are shaped by the FSA ownership
and contracting and licensing arrangements. This research suggests that there can be
significant operational ramifications for removing ownership rights away from value
creating subsidiaries. Moreover, pure tax havens can have a negative impact on MNC
innovation. When ownership is granted to pure tax haven entities, MNCs are challenged
by not only the difficulties of managing innovation, but also the negative incentive effect
from profits going to a non-value generating unit. Subsidiary ties within the firm are
126
important for knowledge sharing and innovation (O’Donnell, 2000). Pure tax haven
entities may cause a breakdown the linkages between organizational units.
This paper makes several important contributions to the literature on the MNC,
subsidiary entrepreneurship and innovation. First, it contributes to the literature on the
multinational firm and innovation by examining the role of the internal ownership of
FSAs on affiliate innovation, which has thus far been unexamined. The allocation of
ownership of FSAs has strategic ramifications on the future development of MNC FSAs.
The results provide strong support that subsidiary-level ownership of FSAs can be used
to enhance value.
Second, I contribute to the subsidiary entrepreneurship literature by introducing
economic ownership of FSAs as a means of establishing “entrepreneurial” subsidiaries
within the firm. The economic owners are the “entrepreneurs of the firm” in that they
bear the risks and reap the rewards from the MNC’s activities. In contrast to the
centralized view of parent decision making and control, a number of researchers view the
multinational firm as an interorganizational network (e.g. Ghoshal and Bartlett, 1990;
Hedlund, 1986; Gupta and Govindarajan, 1991). The network perspective assumes that
subsidiaries hold strategic roles in the development and maintenance of FSAs. Along
these lines, researchers studying subsidiary world mandates suggest that subsidiaries can
be granted global responsibility for a product line (e.g. Birkinshaw, 1996; Rugman,
1981). This study shares the view that subsidiaries can hold important roles in the
development and maintenance of FSAs and in bearing global responsibilities. However,
the activities of the FSA owner encompass contracting and licensing other entities within
the firm. In contracting others subsidiaries to perform activities on their behalf, the FSA
127
owners insulate the contracted subsidiaries from risk. The FSA users do not gain from
exerting extra efforts or bear the consequence of their actions. This research suggests
that some subsidiaries may have power over others. That is, the FSA owner can
influence the mandates that are gained and lost by the FSA users within the firm.
There are several limitations to this research. First, this study is limited in that it
analyzes technological innovations using patents. Certain types of innovations are more
suitable to patenting than others, which leads to industry and firm differences in patenting
behavior. Additionally, patenting activities can be driven by firm strategies for
intellectual property management. I attempt to control for industry-level differences in
patenting behaviors using industry fixed effects. Second, this research studies
technological innovations only. Future research using measures of different types of
innovations can enhance our understanding of the effects of FSA ownership on
innovation. Third, the analysis of ownership change is limited by the small number of
observations. Care should be taken in interpreting the results until further analysis is
done.
There are a number of potential areas for future research. First, an examination of
how internal and external contract and licensing arrangements are similar to or different
from each other can enhance our understanding of the theory of the MNC. Second, there
are clear differences in knowledge sharing behaviors for FSA owners and users, as well
as for the FSA ownership structures. Future research exploring the effects of FSA
ownership on knowledge sharing within the firm can provide better insight into how
contracting relationships affect innovation.
128
6. References
Aghion P, Tirole, J. 1997. “Formal and Real Authority in Organization,” Journal of
Political Economy 105: 1-29.
Ahuja, G. 2000. Collaboration Networks, Structural Holes, and Innovation: A
Longitudinal Study. Administrative Science Quarterly, 45(3): 425-455.
Alchian A, Demsetz H. 1972. Production, Information Costs, and Economic
Organization. American Economic Review. 62(5): 772-795.
Alchian A, Demsetz H. 1973. The Property Right Paradigm. The Journal of Economic
History. 33(01): 16-27.
Almeida P, Phene A. 2004. “Subsidiaries and Knowledge Creation: The Influence Of
The MNC And Host Country On Innovation.” Strategic Management Journal, 25:
847-864.
Ambrosini V, Bowman C. 2001. Tacit Knowledge: Some Suggestions for
Operationalization. Journal of Management Studies, 38(6): 811-829.
Amit R, Schoemaker PJ. 1993. Strategic Assets and Organizational Rent. Strategic
Management Journal, 14(1): 33-46.
Anand J, Kogut, B. 1997. Technological Capabilities of Countries, Firm Rivalry and
Foreign Direct Investment. Journal of International Business Studies, 28(3): 445-
465.
Anderson E, Schmittlein D. 1984. Integration of the Sales Force: An Empirical
Examination. The RAND Journal of Economics 15(3): 385-395.
Andersen T, Denrell J, Bettis, R. 2007. “Strategic Responsiveness and Bowman’s Risk-
Return Paradox,” Strategic Management Journal, 28: 407–429.
Andersson U, Bjorkman I. and Forsgren, M. 2005. “Managing Subsidiary Knowledge
Creation: The Effect of Control Mechanisms on Subsidiary Local
Embeddedness,” International Business Review, 14: 521–538.
Asher C, Mahoney J, and Mahoney J. 2005. “Towards a Property Rights Foundation for
a Stakeholder Theory of the Firm,” Journal of Management and Governance, 9:
5–32.
Banbury C, Mitchell W. 1995. “The Effect of Introducing Important Incremental
Innovations on Market Share and Business Survival.” Strategic Management
Journal, 16: 161-182.
129
Bartlett C, Ghoshal S. 1989. Managing Across Borders: The Transnational Solution.
Boston. MA: Harvard Business School Press.
Barzel Y. 1997. Economic Analysis of Property Rights, 2nd ed. Cambridge University
Press: Cambridge.
Baysinger B, Hoskisson, R. 1989. “Diversification Strategy and R&D Intensity in
Multiproduct Firms,” The Academy of Management Journal, 32(2): 310-332.
Belderbos R. 2003. Entry Mode, Organizational Learning, and R&D in Foreign
Subsidiaries: Evidence from Japanese Firms. Strategic Management Journal,
24(3): 235-359.
Belderbos R, Carree M, Lokshin B. 2004. Cooperative R&D and Firm Performance.
Research Policy, 33: 1477–1492.
Bergin, T. 2012. Special Report: How Starbucks Avoids UK Taxes. Reuters. October
15, 2012.
Birkinshaw J. 1996. How Multinational Subsidiary Mandates are Gained and Lost.
Journal of International Business Studies, 27(3): 467-495.
Birkinshaw J, Hood N. 1998. Multinational Subsidiary Development: Capability
Evolution and Charter Change in Foreign-Owned Subsidiary Companies.
Academy of Management Review, 23: 773-795.
Blundell R, Griffith R, Van Reenen J. 1999. Market Share, Market, Value and
Innovation in a Panel of British Manufacturing Firms. Review Economic Study,
66(3): 529-554.
Boddewyn, J. 1988. “Political Aspects of MNE Theory,” Journal of International
Business Studies, Vol.19(3): 341-363.
Boddewyn J, Brewer T. 1994. International-Business Political Behavior: New
Theoretical Directions, The Academy of Management Review, 19(1): 119-143.
Bowen H, Wiersema M. 2005. Foreign-Based Competition and Corporate
Diversification Strategy. Strategic Management Journal, 26(12): 1153-1171.
Bouquet C, Birkinshaw J. 2008. Weight Versus Voice: How Foreign Subsidiaries Gain
Attention from Corporate Headquarters. Academy of Management Journal,
51(3): 577-601.
130
Brusco S, Panunzi F. 2005. Reallocation of Corporate Resources and Managerial
Incentives in Internal Capital Markets. European Economic Review, 49(3): 659-
681.
Buckley P, Casson M. 1976. The Future of the Multinational Enterprise. London:
Holmes and Meier.
Buckley P, Casson M. 2009. The Internalisation Theory of the Multinational Enterprise:
A Review of the Progress of a Research Agenda After 30 Years Journal of
International Business Studies, 40: 1563–1580.
Buckley P, Hashai N. 2009. Formalizing Internationalization in the Eclectic Paradigm.
Journal of International Business Studies, 40(1): 58-70.
Burk D, Donnell B. 2007. The Goldilocks Hypothesis Balancing Intellectual Property
Rights at the Boundary of the Firm. The University of Illinois Law Review, 2007:
575-636.
Birkinshaw J, Hood N, Jonsson S. 1998. Building Firm-Specific Advantages in
Multinational Corporations: The Role of Subsidiary Initiative. Strategic
Management Journal, 19: 221-241.
Cantwell J, 1995. The Globalization of Technology: What Remains of the Product Cycle
Model? Cambridge Journal of Economics, 19, 155–174.
Cantwell, J. 1987. The Reorganization of European Industries After Integration:
Selected Evidence on the Role of Transnational Enterprise Activities. Journal of
Common Market Studies, 26: 127-151.
Caves, R. 1971. Industrial Corporations: The Industrial Economics of Foreign
Investment. Economica, 38: 1–27.
Coase, R. 1937. The Nature of the Firm. Economica, 4: 386-405.
Cohen W, Levin R, Mowery D. 1987. Firm Size and R&D Intensity: A Re-examination.
Journal of Industrial Economics, 35: 543-63.
Cohen W, Levinthal D. 1989. Innovation and Learning: The Two Faces of R&D.
Economic Journal, 99: 569-596.
Cohen W, Levinthal D. 1990. Absorptive Capacity: A New Perspective on Learning
and Innovation, Administrative Science Quarterly, 35(1): 128-152.
Cohen, W. 2010. Fifty Years of Empirical Studies of Innovative Activity and
Performance. Handbook of the Economics of Innovation, 1: 129-213.
131
Cottrell T, Nault B. 2004. “Product Variety and Firm Survival in the Microcomputer
Software Industry,” Strategic Management Journal, 25(10): 1005–1025.
Cuervo-Cazurra A, Maloney M, Manrakhan S. 2007. Causes of the Difficulties in
Internationalization. Journal of International Business Studies, 38: 709-725.
Cuervo-Cazurra A, Un CA. 2007. Regional Economic Integration and R&D Investment.
Research Policy, 36: 227-246.
Dharmapala D, and Hines J. 2009. Which Countries Become Tax Havens? Journal of
Public Economics, 93(9): 1058-1068.
Damanpour F. 1991. Organizational Innovation: A Meta-Analysis of Effects of
Determinants and Moderates. Academy of Management Journal. 34(3), 555-590.
Das T, Teng B. 2001. A Risk Perception Model of Alliance Structuring. Journal of
International Management, 7:1-29.
Debreu G. 1960. Individual Choice Behavior – A Theoretical Analysis – Luce RD.
American Economic Review, 50: 186–188.
Dischinger M, Riedel N. 2010. The Role of Headquarters Firms in Multinational Profit
Shifting Strategies, Discussion Paper.
Dosi G. 1988. “Sources, Procedures, and Microeconomic Effects of Innovation,”
Journal of Economic Literature, 26(3): 1120-1171.
Dow J, Endersby J. 2004. Multinomial Probit and Multinomial Logit: A Comparison of
Choice Models for Voting Research. Electoral Studies, 23: 107–122.
Driffield N, Love J. 2003. Foreign Direct Investment, Technology Sourcing and
Reverse Spillovers. The Manchester School, 71(6): 659–672.
Drucker J. 2010. Google 2.4% Rate Shows How $60 Billion in Lost to Tax Loopholes.
http://www.bloomberg.com/news.
Drucker P. 1995. Multinomial Probit and Multinomial Logit: A Comparison of Choice
Models for Voting Research. Electoral Studies, 23: 107–122.
Drucker P. 2010. The Frontiers of Management. Harvard Business Press.
Dunning J. 1977. Trade Location of Economic Activity and the Multinational
Enterprise. A Search for an Eclectic Approach, The International Allocation of
Economic Activity.
132
Dunning J. 1980. Toward an Eclectic Theory of International Production: Some
Empirical Tests. Journal of International Business Studies, 11(1): 9–31.
Dunning J, Narula R, 1995. The R&D Activities of Foreign Firms in the United States.
International Studies of Management & Organization, 25(1-2): 29-73.
Elfenbein D, Lerner J. 2003. Ownership and Control Rights in Internet Portal Alliances,
1995-1999. The RAND Journal of Economics, 34(2): 356-369.
Ethier W. 1986. The Multinational Firm. The Quarterly Journal of Economics, 101(4):
805-834.
E&Y. 2013. 2013 Transfer Pricing Global Reference Guide.
Fama E. 1980. Agency Problems and the Theory of the Firm. The Journal of Political
Economy, 88(2): 288-307.
Fama E, Jensen M. 1983. Separation of Ownership and Control. Journal of Law and
Economics. 26(2): 301-325.
Feenstra RC, Hanson GH. 2005. Ownership and Control in Outsourcing to China:
Estimating the Property-Rights Theory of the Firm. The Quarterly Journal of
Economics, 120(2): 729-761.
Feinberg S, Keane M. 2006. Accounting for the Growth of MNC-based Trade using a
Structural Model of U.S. MNCs. American Economic Review, 96(5): 1515-1558.
Feldman D. 1976. The Role of Initiation Activities in Socialization. Human Relations
SAGE Journal, 30(11): 977-990.
Fey C, Furu P. 2008. Top Management Incentive Compensation and Knowledge
Sharing in Multinational Corporations. Strategic Management Journal, 29(12):
1301-1323.
Florida R, Kenney M. 1994. The Globalization of Japanese R&D: The Economic
Geography of Japanese R&D Investment in the United States. Economic
Geography, 70(4): 344-369.
Foss K, Foss N. 1998. The Knowledge-Based Approach: An Organizational Economics
Perspective. Working Paper, Department of Industrial Economics and Strategy,
Copenhagen Business School.
Foss K, Foss N. 2001. Assets, Attributes, and Ownership. International Journal of
Economics of Business. 8(1): 19-37.
133
Foss, N. 2003. Selective Intervention and Internal Hybrids: Interpreting and Learning
From the Rise and Decline of the Oticon Spaghetti Organization. Organization
Science, 14(3): 331-349.
Frost, T. 2001. The Geographic Sources Of Foreign Subsidiaries’ Innovations.
Strategic Management Journal, 22: 101-123.
Frost T, Birkinshaw J, Ensign P. 2002. Centers of Excellence in Multinational
Corporations. Strategic Management Journal, 23: 997-1118.
Frost T, Zhou C. 2005. R&D Co-Practice and 'Reverse' Knowledge Integration in
Multinational Firms. Journal of International Business Studies, 36(6): 676-687.
Galbraith, C. 1990. Transferring Core Manufacturing Technologies in High-Technology
Firms. California Management Review, 32(4): 56-70.
Garcia-Vega M. 2006. Does Technological Diversification Promote Innovation?: An
Empirical Analysis for European Firms. Research Policy, 35: 230-246.
Gautier A, Heider F. 2009. The Benefit and Cost of Winner-Picking: Redistribution
versus Incentives. Journal of Institutional and Theoretical Economics, 165(4):
622-649.
Gertner, Powers, Sharfstein. 2002. Learning about Internal Capital Markets from
Corporate Spin-offs. Journal of Finance, 57(6): 2479-2506
Ghoshal S, Bartlett C. 1990. The Multinational Corporation as an Interorganizational
Network. Academy of Management Review, 15: 603-625.
Ghoshal S, Bartlett CA. 1990. The Multinational Corporation as an Interorganizational
Network. Academy of Management Review, 15: 626-625.
Ghoshal S, Korine H, Szulanski G. 1994. Interunit Communication in Multinational
Corporations. Management Science, 40(1): 96-110.
Grant R. 1996. Toward a Knowledge-Based Theory of the Firm. Strategic Management
Journal, 17: 109-122.
Greene W. 2000. Econometric Analysis (4th ed.). Upper Saddle River, NJ: Prentice-Hall.
Greve H. 2003. A Behavioral Theory of R&D Expenditures and Innovations: Evidence
from Shipbuilding. Academy of Management Journal, 46(6): 685-702.
Grossman S, Hart O. 1986. The Costs and Benefits of Ownership: A Theory of Lateral
and Vertical Integration. Journal of Political Economy, 94: 691-719.
134
Grubert H, Mutti J. 1991. Taxes, Tariffs and Transfer Pricing in Multinational Corporate
Decision Making. The Review of Economics and Statistics, 73(2): 285-293.
Grubert, H. 2003. Intangible Income, Intercompany Transactions, Income Shifting, and
the Choice of Location. National Tax Journal, 221-242.
Grubert H, Mutti J. 2008. The Effect of Taxes on Royalties and the Migration of
Intangible Assets Abroad. CRIW Conference on International Service Flows.
Gupta A, Govindarajan V. 1991. Knowledge Flows and the Structure of Control within
Multinational Corporations. Academy of Management Review, 16(4): 768-792.
Gupta A, Govindarajan V. 2000. Flows within Multinational Corporations. Strategic
Management Journal, 21(4): 473-496
Håkanson L, Nobel. 1993. Foreign Research and Development in Swedish
Multinationals. Research Policy, 22(5-6): 373-396.
Hall, Griliches, Hausman. 1986. Patents and R&D: Is There A Lag? NBER Working
Paper No. 1454 (Also Reprint No. r0775).
Hall B, Jaffe, Trajtenberg M. 2001. Market Value and Patent Citations: A First Look.
NBER Working Paper No. w7741.
Hamilton R, Kashlak R. 1999. National Influences on Multinational Corporation
Control System Selection. Management International Review, 39(2): 167-189.
Hansen M. 1999. The Search-Transfer Problem: The Role of Weak Ties in Sharing
Knowledge Across Organization Subunits. Administrative Science Quarterly, 44:
82-111.
Hart O, Moore J. 1990. Property Rights and the Nature of the Firm. Journal of Political
Economy, 98: 1119-1158.
Hart O. 1995. Corporate Governance: Some Theory and Implications. The Economic
Journal, 105(430): 678-689.
Hart O. 2001. Financial Contracting. NBER Working Paper No. 8285.
Hedlund G. 1986. The Hypermodern MNC-A Heterarchy? Human Resource
Management, 25: 9-25.
Hedlund, G. 1994. A Model of Knowledge Management and the N-form Corporation.
Strategic Management Journal, 15(S-2), 73-90.
135
Helfat C. 1997. Know-how and Asset Complementarity and Dynamic Capability
Accumulation: The Case of R&D. Strategic Management Journal, 18: 339-360.
Helfat C. 2000. The Evolution of Firm Capabilities. Strategic Management Journal, 21:
955-959.
Henderson B. 1979. Henderson on Corporate Strategy. Cambridge, Mass.: Abt Books.
Henderson R. 1993. Underinvestment and Incompetence as Responses to Radical
Innovation: Evidence from the Photolithographic Alignment Equipment Industry.
RAND Journal of Economics, 24(2): 248-270.
Henderson, R., Cockburn, I. 1994. Measuring Competence? Exploring Firm Effects in
Pharmaceutical Research. Strategic Management Journal, 15(S-1): 63-84.
Henisz W. 2000. The Institutional Environment for Multinational Investment. Journal
of Law, Economics and Organization, 16(2): 334-364.
Himmelberg C, Hubbard R, Love I. 2002. Investor Protection, Ownership, and the Cost
of Capital.
Hilbe J. 2011. Negative Binomial Regression. Cambridge University Press, New York.
Holmstrom B, Milgrom P. 1991. Multitask Principal-Agent Analyses: Incentive
Contracts, Asset Ownership and Job Design. Journal of Law, Economics &
Organization, 7: 24.
Holmstrom B, Milgrom P. 1994. The Firm as an Incentive System. The American
Economic Review, 84(4): 972-991.
Holmstrom B, Roberts J. 1998. The Boundaries of the Firm Revisited. The Journal of
Economic Perspectives, 12(4): 73-94.
Horstein A, Zhao M. 2011. Corporate Capital Budgeting Decisions and Information
Sharing. Journal of Economics & Management Strategy, 20(4): 1135–1170.
Hoskisson R, Hitt M, Hill C. 1993. Managerial Incentives and Investment in R&D in
Large Multiproduct Firms. Organization Science, 4(2): 325-341.
Hoskisson R, Johnson R. 1992. Research Notes and Communications Corporate
Restructuring and Strategic Change: The Effect on Diversification Strategy and
R&D Intensity. Strategic Management Journal, 13(8): 625–634.
Hymer S. 1960. The International Operations of National Firms: A Study of Direct
Foreign Investment. Ph.D. Dissertation, Massachusetts Institute of Technology
(published by MIT Press, 1976).
136
Inderst R, Laux C. 2005. Incentives in Internal Capital Markets: Capital Constraints,
Competition, and Investment Opportunities. The RAND Journal of Economics
36(1): 215-228.
Jaffe E, Nebenzahl I. 2001. National Image and Competitive Advantage: The Theory
and Practice of Country-of-Origin Effect. Copenhagen Business School Press:
Copenhagen.
Jap S, Robertson D, Hamilton R. 2011. The Dark Side of Rapport: Agent Misbehavior
Face-to-Face and Online. Management Science 57(9): 1610-1622.
Jensen M, Meckling W. 1976. Theory of the Firm: Managerial Behavior, Agency Costs
and Ownership Structure. Journal of Financial Economics, 3(4): 305-360.
Jensen M. 1986. Agency Costs of Free Cash Flows, Corporate Finance and Takeovers.
American Economic Review, 76(2): 323-329.
Kaplan S, Stromberg P. 2003. Financial Contracting Theory Meets the Real World: An
Empirical Analysis of Venture Capital Contracts. Review of Economic Studies,
70: 281-315.
Katila R, Ahuja G. 2002. Something Old, Something New: A Longitudinal Study of
Search Behavior and New Product Introduction. Academy of Management
Journal, 45: 1183-1194.
Kim J, Mahoney J. 2002. Resource-Based and Property Rights Perspectives on Value
Creation: The Case of Oil Field Unitization. Managerial and Decision
Economics, 23(4/5): 225-245.
Kim J, Mahoney J. 2005. Property Rights Theory, Transaction Costs Theory, and
Agency Theory: An Organizational Economics Approach to Strategic
Management. Managerial and Decision Economics, 26(4): 223-242.
Kitch E. 1980. The Law and Economics of Rights in Valuable Information. The
Journal of Legal Studies, 9(4): 683-723.
Klein B. 1983. Contracting Costs and Residual Claims: Then Separation of Ownership
and Control. The Journal of Law and Economics, 26: 367-374.
Kogut B, Chang S. 1991. Technological Capabilities and Japanese Foreign Direct
Investment in the United States. The Review of Economics and Statistics, 73(3):
401-413.
Kogut B, Zander U. 1992. Knowledge of the Firm, Combinative Capabilities and the
Replication of Technology. Organization Science, 3: 383–397.
137
Kogut B, Zander U. 1993. Knowledge of the Firm and the Evolutionary Theory of the
Multinational Corporation. Journal of International Business Studies, 625-645.
Kuemmerle, W. 1996. Foreign Direct Investment in Industrial Research in the
Pharmaceutical and Electronics Industries – Results from a Survey of
Multinational Firms. Research Policy, 28(2-3): 179-193.
Kuemmerle W. 1997. Building Effective R&D Capabilities Abroad. Harvard Business
Review, 75: 61-70.
Kuemmerle W. 1999. The Drivers of Foreign Direct Investment in Research and
Development: An Empirical Investigation. Journal of International Business
Studies, 30(1): 1-24.
Lafontaine F, Slade M. 2007. Vertical Integration and Firm Boundaries: The Evidence.
Journal of Economic Literature, 45(3): 629-685.
Lambert R, Larcker D, Weigelt K. 1993. The Structure of Organizational Incentives.
Administrative Science Quarterly, 438-461.
Lee, O’Neill. 2003. Ownership Structures and R&D Investments of U.S. and Japanese
Firms: Agency and Stewardship Perspectives. The Academy of Management
Journal, 46: 212-225.
Leiponen A. 2008. Control of Intellectual Assets in Client Relationships: Implications
for Innovation. Strategic Management Journal, 29(13): 1371-1394.
Lerner J, Malmendier U. 2010. Contractibility and the Design of Research Agreements.
American Economic Review, 100(1): 214–246.
Levin R, Cohen W, Mowery D. 1987. R&D Appropriability, Opportunity, and Market
Structure: New Evidence on Some Schumpeterian Hypothesis. The American
Economic Review, 75(2): 20-24.
Levin C, McCain J. 2013. Offshore Profit Shifting and the U.S. Tax Code - Part 2
(Apple Inc.). Memorandum to Members of the Permanent Subcommittee on
Investigations. May 21, 2013.
Lippman S, Rumelt R. 1982. Uncertain Imitability: An Analysis of Interfirm
Differences in Efficiency under Competition. The Bell Journal of Economics,
13(2): 418-438.
Madhok A. 2002. Reassessing the Fundamentals and Beyond: Ronald Coase, the
Transaction Cost and Resource-Based Theories of the Firm and the Institutional
Structure of Production. Strategic Management Journal, 23: 535–550.
138
Maksimovic V, Phillips G. 2002. Do Conglomerate Firms Allocate Resources
Inefficiently Across Industries? Journal of Finance, 57(2), 721-767.
Martin X, Salomon R. 2003. Knowledge Transfer Capacity and its Implications for the
Theory of the Multinational Corporation. Journal of International Business
Studies, 34(4): 356-373.
Maskin E, Tirole J. 1999. Two Remarks on the Property-Rights Literature. The Review
of Economic Studies, 66(1): 139-149.
Masten S. 1984. The Organization of Production: Evidence from the Aerospace
Industry, Journal of Law and Economics 27: 403-417.
Matouschek N. 2004. Ex Post Inefficiencies in a Property Rights Theory of the Firm.
Journal of Law, Economics and Organization, 20(1): 125-147.
McFadyen M, Cannella A. 2004. Social Capital and Knowledge Creation: Diminishing
Returns of the Number and Strength of Exchange. The Academy of Management
Journal, 47(5): 735-746.
Morck R, Yeung B. 1991. Why Investors Value Multinationality. The Journal of
Business, 64(2): 165-187.
Mudambi R, Mudambi S, Navarra P. 2007. Global Innovation in MNCs: The Effects of
Subsidiary Self-Determination and Teamwork. Journal of Product Innovation
Management, 24(5): 442-455.
Nelson R, Winter G. 1982. An Evolutionary Theory of Economic Change. Cambridge:
Belknap.
Nerkar A, Roberts P. 2004. Technological and Product-Market Experience and the
Success of New Product Introductions in the Pharmaceutical Industry. Strategic
Management Journal, 25(8/9): 779-799.
Nobel R, Birkinshaw J. 1998. Innovation in Multinational Corporations: Control and
Communication Patterns in International R&D Operations. Strategic
Management Journal, 19: 479-496.
Nohria N, Ghoshal S. 1997. The Differentiated Network: Organizing Multinational
Corporations for Value Creation. San Francisco: Jossey-Bass.
OECD 2000. Towards Global Tax Cooperation: Progress in Identifying and Eliminating
Harmful Tax Practices. Paris: OECD.
139
OECD. 2009. OECD Science, Technology and R&D Statistics Online Database. Paris:
Organization for Economic Cooperation and Development.
OECD. 2013. Revised Discussion Draft on Transfer Pricing Aspects of Intangibles.
Paris: Organization for Economic Cooperation and Development.
Osterloh M, Frey B. 2000. Motivation, knowledge Transfer, and Organizational Forms.
Organization Science, 11: 538-550.
Overesch M, Schreiber U. 2008. R&D Intensities, International Profit Shifting, and
Investment Decisions. Discussion Paper.
Palepu K. 1985. Diversification Strategy, Profit Performance and the Entropy Measure.
Strategic Management Journal, 6(3): 239-255.
Park S, Russo M. 1996. When Competition Eclipses Cooperation: An Event History
Analysis of Joint Venture Failure. Management Science, 42(6): 875-890.
Pearce R, Papanastassiou M. 1999. Overseas R&D and the Strategic Evolution of
MNEs: Evidence from Laboratories in the UK. Research Policy, 28: 23–41.
Penrose E. 1959. The Theory of the Growth of the Firm. London: Blackwells.
Peteraf M. 1993. The Cornerstones of Competitive Advantage: A Resource Based
View. Strategic Management Journal, 14: 179-191.
Phene A, Almeida P. 2008. Innovation in Multinational Subsidiaries: The Role of
Knowledge Assimilation and Subsidiary Capabilities. Journal of International
Business Studies, 39(5): 901-919.
Pisano G. 1990. The R&D Boundaries of the Firm: An Empirical Analysis.
Administrative Science Quarterly, 35: 153-76.
Pisano G. 1991. The Governance of Innovation: Vertical Integration and Collaborative
Arrangements in the Biotechnology Industry. Research Policy, 20: 237-49.
Rajan R, Zingales L. 1998. Power in a Theory of a Firm. Quarterly Journal of
Economics, 13(2): 387-432.
Rajan R, Servaes H, Zingales L. 2000. The Cost of Diversity: The Diversification
Discount and Ineffective Investment. The Journal of Finance, 55(1): 35-80.
Rajan R, Zingales L. 2000. The Governance of the New Enterprise. Corporate
Governance, Theoretical & Empirical Perspectives, 10: 1-42.
140
Rosenbaum P, Rubin D. 1983. The Central Role of the Propensity Score in
Observational Studies for Causal Effects. Biometrica, 70(1): 41-55.
Rotemberg J, Saloner G. 1994. Benefits of Narrow Business Strategies. The American
Economic Review, 84(5): 1330-1349.
Roth K, O'Donnell S. 1996. Foreign Subsidiary Compensation Strategy: An Agency
Theory Perspective. The Academy of Management Journal, 39(3): 678-703.
Rothaermel F, Hess A. 2007. Building Dynamic Capabilities: Innovation Driven by
Individual-, Firm-, and Network-Level Effects. Organization Science, 18(6):
898-921.
Rugman A. 1980. Internalization as a General Theory of Foreign Direct Investment: A
Re-appraisal of the Literature. Review of World Economics, 116(2): 365-379.
Rugman A. 1981. Inside the Multinationals: The Economics of Internal Markets. New
York: Columbia Press. Reissued by Palgrave Macmillan in 2006 as Inside the
Multinationals, (25th
Anniversary Edition), Basingstoke: Palgrave Macmillan.
Rugman A, Verbeke A. 1992. A note on the transnational solution and the transaction
cost theory of multinational strategic management. Journal of International
Business Studies 23(4): 761-772.
Rugman A, Verbeke A. 2001. Subsidiary-Specify Advantages in Multinational
Enterprises, Strategic Management Journal, 22: 237-250.
Sadowski B, Sadowski-Rasters G. 2006. On the Innovativeness of Foreign Subsidiaries:
Evidence from Companies in The Netherlands. Research Policy, 35(3): 447-462.
Simonin B. 1999. Transfer of Marketing Know-How in International Strategic
Alliances: An Empirical Investigation of the Role and Antecedents of Knowledge
Ambiguity. Journal of International Business Studies, 30(3): 463–90.
Spender J, Grant R. 1996. Knowledge and the Firm: Overview. Strategic Management
Journal, 17: 5-9.
Teece D. 1977. Technology Transfer by Multinational Firms: The Resource Costs of
Transferring Technological Know-how. The Economic Journal, 87: 242–261.
Teece D. 1986. Profiting from Technological Innovation: Implications for Integration,
Collaboration, Licensing and Public Policy. Research Policy, 15(6): 285-305.
141
Teece D. 1992. Competition, Cooperation, and Innovation: Organizational
Arrangements for Regimes of Rapid Technological Progress. Journal of
Economic Behavior and Organization, 18: 1-25.
Tirole J. 1999. Incomplete Contracts: Where Do We Stand? Econometrica. 67: 741-781.
Thompson M. 2012. U.K. Targets Google, Amazon and Starbucks on Taxes. CNN
Money. December 3, 2012.
Trajtenberg M. 1990. A Penny for Your Quotes: Patent Citations and the Value of
Innovations. The RAND Journal of Economics, 21(1): 172-187.
Tsai W, Ghoshal S. 1998. Social Capital and Value Creation: The Role of Intrafirm
Networks. Academy of Management Journal, 41: 464–476.
Tsay A. 1999. The Quantity Flexibility Contract and Supplier-Customer Incentives.
Management Science, 45(10): 1339 –1358.
Van Maanen J, Eastin J, Schein E. 1977. Toward a Theory of Organizational
Socialization. Sloan School of Management, 1: 84-89.
Von Hippel E. 1988. Sources of Innovation. Oxford University Press: New York.
Von Zedtwitz M, Gassmann O. 2002. Market Versus Technology Drive in R&D
Internationalization: Four Different Patterns of Managing Research and
Development. Research Policy, 31: 569–588.
Wang H, He J, Mahoney J. 2009. Firm-Specific Knowledge Resources and Competitive
Advantage: The Roles of Economic- and Relationship-Based Employee
Governance Mechanisms. Strategic Management Journal, 30(12): 1265-1285.
Whinston M. 2001. The Scope of the Firm: New Empirical Directions Assessing the
Property Rights and Transaction-Cost Theories of Firm Scope. AEA Papers and
Proceedings, 91(2): 184-188.
Zander U, Kogut B. 1995. Knowledge and the Speed of the Transfer and Imitation of
Organizational Capabilities: An Empirical Test. Organizational Science, 6(1):
76-92.
Zander I. 1997. Technological Diversification in the Multinational Corporation–
Historical Evolution and Future Prospects. Research Policy, 26: 209–227.
Zhao M. 2006. Conducting R&D in Countries with Weak Intellectual Property Rights
Protection. Management Science, 52(8): 1185-1199.
Zingales L. 2000. In Search of New Foundations. The Journal of Finance 55: 1623-1653.
142
Curriculum Vitae
Catherine Magelssen
Rutgers University, Department of Management and Global Business
1 Washington Park, Newark, NJ 07102
1982 Born in Turlock, California.
2000 High School Diploma, Trinity Christian School, Sacramento, California.
2002 A.A. Modesto Junior College, Modesto, California.
2004 B.S. Business Administration, Haas School of Business, University of
California, Berkeley.
2004-2008 Senior Associate. The Ballentine Barbera Group, A CRA International
Company, Palo Alto/Pleasanton, California.
2014 PhD in Management, Rutgers Business School, Newark, New Jersey.
Publications and Proceedings
2008 Curd, R., Hart, R., and Magelssen, C. “Modeling Risk,” Euromoney ITR
Intellectual Property Guide.
2010 Feinberg, S., Magelssen, C., and Smith, M. “Early Adopters and Radical
Regulatory Reform in India: First in Line for Trade Protection,” Academy
of Management Best Paper Proceedings.
2012 Damanpour, F., and Magelssen, C. “Adoption Cycle of Managerial
Innovation: Effects of Innovation Social Status and Transaction Costs,”
Academy of Management Best Paper Proceedings.
2012 Damanpour, F., Chiu, H., and Magelssen, C. “Initiation, Implementation,
and Complexity of Managerial Innovation,” Handbook of Organizational
and Managerial Innovation, T. S. Pitsis, A. Simpson, and E. Dehlin (Ed.),
Edward-Elgar Press.